Other functions

Documentation for the “other” functions category

The function documentation below is generated from the system.functions system table.

FQDN

Introduced in: v20.1.0

Returns the fully qualified domain name of the ClickHouse server.

Syntax

FQDN()

Aliases: fullHostName

Arguments

• None.

Returned value

Returns the fully qualified domain name of the ClickHouse server. String

Examples

Usage example

Query

SELECT fqdn()

Response

┌─FQDN()──────────────────────────┐
│ clickhouse.us-east-2.internal │
└─────────────────────────────────┘

MACNumToString

Introduced in: v1.1.0

Interprets a UInt64 number as a MAC address in big endian format. Returns the corresponding MAC address in format AA:BB:CC:DD:EE:FF (colon-separated numbers in hexadecimal form) as string.

Syntax

MACNumToString(num)

Arguments

• num — UInt64 number. UInt64

Returned value

Returns a MAC address in format AA:BB:CC:DD:EE:FF. String

Examples

Usage example

Query

SELECT MACNumToString(149809441867716) AS mac_address;

Response

┌─mac_address───────┐
│ 88:00:11:22:33:44 │
└───────────────────┘

MACStringToNum

Introduced in: v1.1.0

The inverse function of MACNumToString. If the MAC address has an invalid format, it returns 0.

Syntax

MACStringToNum(s)

Arguments

• s — MAC address string. String

Returned value

Returns a UInt64 number. UInt64

Examples

Usage example

Query

SELECT MACStringToNum('01:02:03:04:05:06') AS mac_numeric;

Response

1108152157446

MACStringToOUI

Introduced in: v1.1.0

Given a MAC address in format AA:BB:CC:DD:EE:FF (colon-separated numbers in hexadecimal form), returns the first three octets as a UInt64 number. If the MAC address has an invalid format, it returns 0.

Syntax

MACStringToOUI(s)

Arguments

• s — MAC address string. String

Returned value

First three octets as UInt64 number. UInt64

Examples

Usage example

Query

SELECT MACStringToOUI('00:50:56:12:34:56') AS oui;

Response

20566

authenticatedUser

Introduced in: v25.11.0

If the session user has been switched using the EXECUTE AS command, this function returns the name of the original user that was used for authentication and creating the session. Alias: authUser()

Syntax

authenticatedUser()

Aliases: authUser

Arguments

• None.

Returned value

The name of the authenticated user. String

Examples

Usage example

Query

EXECUTE as u1;
            SELECT currentUser(), authenticatedUser();

Response

┌─currentUser()─┬─authenticatedUser()─┐
│ u1            │ default             │
└───────────────┴─────────────────────┘

bar

Introduced in: v1.1.0

Builds a bar chart. Draws a band with width proportional to (x - min) and equal to width characters when x = max. The band is drawn with accuracy to one eighth of a symbol.

Syntax

bar(x, min, max[, width])

Arguments

• x — Size to display. (U)Int* or Float* or Decimal
• min — The minimum value. (U)Int* or Float* or Decimal
• max — The maximum value. (U)Int* or Float* or Decimal
• width — Optional. The width of the bar in characters. The default is 80. const (U)Int* or const Float* or const Decimal

Returned value

Returns a unicode-art bar string. String

Examples

Usage example

Query

SELECT
toHour(EventTime) AS h,
count() AS c,
bar(c, 0, 600000, 20) AS bar
FROM test.hits
GROUP BY h
ORDER BY h ASC

Response

┌──h─┬──────c─┬─bar────────────────┐
│  0 │ 292907 │ █████████▋         │
│  1 │ 180563 │ ██████             │
│  2 │ 114861 │ ███▋               │
│  3 │  85069 │ ██▋                │
│  4 │  68543 │ ██▎                │
│  5 │  78116 │ ██▌                │
│  6 │ 113474 │ ███▋               │
│  7 │ 170678 │ █████▋             │
│  8 │ 278380 │ █████████▎         │
│  9 │ 391053 │ █████████████      │
│ 10 │ 457681 │ ███████████████▎   │
│ 11 │ 493667 │ ████████████████▍  │
│ 12 │ 509641 │ ████████████████▊  │
│ 13 │ 522947 │ █████████████████▍ │
│ 14 │ 539954 │ █████████████████▊ │
│ 15 │ 528460 │ █████████████████▌ │
│ 16 │ 539201 │ █████████████████▊ │
│ 17 │ 523539 │ █████████████████▍ │
│ 18 │ 506467 │ ████████████████▊  │
│ 19 │ 520915 │ █████████████████▎ │
│ 20 │ 521665 │ █████████████████▍ │
│ 21 │ 542078 │ ██████████████████ │
│ 22 │ 493642 │ ████████████████▍  │
│ 23 │ 400397 │ █████████████▎     │
└────┴────────┴────────────────────┘

blockNumber

Introduced in: v1.1.0

Returns a monotonically increasing sequence number of the block containing the row. The returned block number is updated on a best-effort basis, i.e. it may not be fully accurate.

Syntax

blockNumber()

Arguments

• None.

Returned value

Sequence number of the data block where the row is located. UInt64

Examples

Basic usage

Query

SELECT blockNumber()
FROM
(
    SELECT *
    FROM system.numbers
    LIMIT 10
) SETTINGS max_block_size = 2

Response

┌─blockNumber()─┐
│             7 │
│             7 │
└───────────────┘
┌─blockNumber()─┐
│             8 │
│             8 │
└───────────────┘
┌─blockNumber()─┐
│             9 │
│             9 │
└───────────────┘
┌─blockNumber()─┐
│            10 │
│            10 │
└───────────────┘
┌─blockNumber()─┐
│            11 │
│            11 │
└───────────────┘

blockSerializedSize

Introduced in: v20.3.0

Returns the uncompressed size in bytes of a block of values on disk.

Syntax

blockSerializedSize(x1[, x2[, ...]])

Arguments

• x1[, x2, ...] — Any number of values for which to get the uncompressed size of the block. Any

Returned value

Returns the number of bytes that will be written to disk for a block of values without compression. UInt64

Examples

Usage example

Query

SELECT blockSerializedSize(maxState(1)) AS x;

Response

┌─x─┐
│ 2 │
└───┘

blockSize

Introduced in: v1.1.0

In ClickHouse, queries are processed in blocks (chunks). This function returns the size (row count) of the block the function is called on.

Syntax

blockSize()

Arguments

• None.

Returned value

Returns the number of rows in the current block. UInt64

Examples

Usage example

Query

SELECT blockSize()
FROM system.numbers LIMIT 5

Response

┌─blockSize()─┐
│           5 │
│           5 │
│           5 │
│           5 │
│           5 │
└─────────────┘

buildId

Introduced in: v20.5.0

Returns the build ID generated by a compiler for the running ClickHouse server binary. If executed in the context of a distributed table, this function generates a normal column with values relevant to each shard. Otherwise it produces a constant value.

Syntax

buildId()

Arguments

• None.

Returned value

Returns the build ID. String

Examples

Usage example

Query

SELECT buildId()

Response

┌─buildId()────────────────────────────────┐
│ AB668BEF095FAA6BD26537F197AC2AF48A927FB4 │
└──────────────────────────────────────────┘

byteSize

Introduced in: v21.1.0

Returns an estimation of the uncompressed byte size of its arguments in memory. For String arguments, the function returns the string length + 8 (length). If the function has multiple arguments, the function accumulates their byte sizes.

Syntax

byteSize(arg1[, arg2, ...])

Arguments

• arg1[, arg2, ...] — Values of any data type for which to estimate the uncompressed byte size. Any

Returned value

Returns an estimation of the byte size of the arguments in memory. UInt64

Examples

Usage example

Query

SELECT byteSize('string')

Response

┌─byteSize('string')─┐
│                 15 │
└────────────────────┘

Multiple arguments

Query

SELECT byteSize(NULL, 1, 0.3, '')

Response

┌─byteSize(NULL, 1, 0.3, '')─┐
│                         19 │
└────────────────────────────┘

catboostEvaluate

Introduced in: v22.9.0

Evaluate an external catboost model. CatBoost is an open-source gradient boosting library developed by Yandex for machine learning. Accepts a path to a catboost model and model arguments (features).

Prerequisites

1. Build the catboost evaluation library

Before evaluating catboost models, the libcatboostmodel.<so|dylib> library must be made available. See CatBoost documentation how to compile it.

Next, specify the path to libcatboostmodel.<so|dylib> in the clickhouse configuration:

<clickhouse>
...
    <catboost_lib_path>/path/to/libcatboostmodel.so</catboost_lib_path>
...
</clickhouse>

For security and isolation reasons, the model evaluation does not run in the server process but in the clickhouse-library-bridge process. At the first execution of catboostEvaluate(), the server starts the library bridge process if it is not running already. Both processes communicate using a HTTP interface. By default, port 9012 is used. A different port can be specified as follows - this is useful if port 9012 is already assigned to a different service.

<library_bridge>
    <port>9019</port>
</library_bridge>

2. Train a catboost model using libcatboost

See Training and applying models for how to train catboost models from a training data set.

Syntax

catboostEvaluate(path_to_model, feature_1[, feature_2, ..., feature_n])

Arguments

• path_to_model — Path to catboost model. const String
• feature — One or more model features/arguments. Float*

Returned value

Returns the model evaluation result. Float64

Examples

catboostEvaluate

Query

SELECT catboostEvaluate('/root/occupy.bin', Temperature, Humidity, Light, CO2, HumidityRatio) AS prediction FROM occupancy LIMIT 1

Response

4.695691092573497

colorOKLABToSRGB

Introduced in: v26.2.0

Converts a color from the OKLab perceptual color space to the sRGB color space.

The input color is specified in the OKLab color space. If the input values are outside the typical OKLab ranges, the result is implementation-defined.

OKLab uses three components:

• L: perceptual lightness (typically in the range [0..1])
• a: green-red opponent axis
• b: blue-yellow opponent axis

The a and b components are theoretically unbounded, but in practice are between -0.4 and 0.4. OKLab is designed to be perceptually uniform while remaining inexpensive to compute.

The conversion is intended to be the inverse of colorSRGBToOKLAB and consists of the following stages:

1. Conversion from OKLab to linear sRGB.
2. Conversion from linear sRGB to gamma-encoded sRGB.

The optional gamma argument specifies the exponent used when converting from linear sRGB to gamma-encoded RGB values. If not specified, a default gamma value is used for consistency with colorSRGBToOKLAB.

For more information about the OKLab color space and its relationship to sRGB, see https://developer.mozilla.org/en-US/docs/Web/CSS/Reference/Values/color_value/oklab .

Syntax

colorOKLABToSRGB(tuple [, gamma])

Arguments

• tuple — A tuple of three numeric values L, a, b, where L is in the range [0...1]. Tuple(Float64, Float64, Float64)
• gamma — Optional. The exponent that is used to transform linear sRGB back to sRGB by applying (x ^ (1 / gamma)) * 255 for each channel x. Defaults to 2.2. Float64

Returned value

Returns a tuple (R, G, B) representing sRGB color values. Tuple(Float64, Float64, Float64)

Examples

Convert OKLAB to sRGB (Float)

Query

SELECT colorOKLABToSRGB((0.4466, 0.0991, 0.44)) AS rgb;

Response

┌─rgb──────────────────────┐
│ (198.07056923258935,0,0) │
└──────────────────────────┘

Convert OKLAB to sRGB (UInt8)

Query

WITH colorOKLABToSRGB((0.7, 0.1, 0.54)) AS t
SELECT tuple(toUInt8(t.1), toUInt8(t.2), toUInt8(t.3)) AS RGB;

Response

┌─RGB──────────┐
│ (255,0,0)    │
└──────────────┘

colorOKLCHToSRGB

Introduced in: v25.7.0

Converts a colour from the OKLCH perceptual colour space to the familiar sRGB colour space.

If L is outside the range [0...1], C is negative, or H is outside the range [0...360], the result is implementation-defined.

OKLCH is a cylindrical version of the OKLab colour space. It’s three coordinates are L (the lightness in the range [0...1]), C (chroma >= 0) and H (hue in degrees from [0...360]). OKLab/OKLCH is designed to be perceptually uniform while remaining cheap to compute.

The conversion is the inverse of colorSRGBToOKLCH:

1. OKLCH to OKLab.
2. OKLab to Linear sRGB
3. Linear sRGB to sRGB

The second argument gamma is used at the last stage.

For references of colors in OKLCH space, and how they correspond to sRGB colors please see https://oklch.com/.

Syntax

colorOKLCHToSRGB(tuple [, gamma])

Arguments

• tuple — A tuple of three numeric values L, C, H, where L is in the range [0...1], C >= 0 and H is in the range [0...360]. Tuple(Float64, Float64, Float64)
• gamma — Optional. The exponent that is used to transform linear sRGB back to sRGB by applying (x ^ (1 / gamma)) * 255 for each channel x. Defaults to 2.2. Float64

Returned value

Returns a tuple (R, G, B) representing sRGB color values. Tuple(Float64, Float64, Float64)

Examples

Convert OKLCH to sRGB

Query

SELECT colorOKLCHToSRGB((0.6, 0.12, 40)) AS rgb;

Response

┌─rgb───────────────────────────────────────────────────────┐
│ (186.02058688365264,100.68677189684993,71.67819977081575) │
└───────────────────────────────────────────────────────────┘

Convert OKLCH to sRGB (UInt8)

Query

WITH colorOKLCHToSRGB((0.6, 0.12, 40)) AS t
SELECT tuple(toUInt8(t.1), toUInt8(t.2), toUInt8(t.3)) AS RGB;

Response

┌─RGB──────────┐
│ (186,100,71) │
└──────────────┘

colorSRGBToOKLAB

Introduced in: v26.2.0

Converts a colour encoded in the sRGB colour space to the perceptually uniform OKLAB colour space.

If any input channel is outside [0...255] or the gamma value is non-positive, the behaviour is implementation-defined.

OKLAB is a perceptually uniform color space. Its three coordinates are L (the lightness in the range [0...1]), a (Green-Red axis) and b (Blue-Yellow axis). OKLab is designed to be perceptually uniform while remaining cheap to compute.

The conversion consists of two stages:

1. sRGB to Linear sRGB
2. Linear sRGB to OKLab

Syntax

colorSRGBToOKLAB(tuple[, gamma])

Arguments

• tuple — Tuple of three values R, G, B in the range [0...255]. Tuple(UInt8, UInt8, UInt8)
• gamma — Optional. Exponent that is used to linearize sRGB by applying (x / 255)^gamma to each channel x. Defaults to 2.2. Float64

Returned value

Returns a tuple (L, a, b) representing the OKLAB color space values. Tuple(Float64, Float64, Float64)

Examples

Convert sRGB to OKLAB

Query

SELECT colorSRGBToOKLAB((128, 64, 32), 2.2) AS lab;

Response

┌─lab──────────────────────────────────────────────────────────┐
│ (0.4436238384931984,0.07266246769242975,0.07500108778529994) │
└──────────────────────────────────────────────────────────────┘

colorSRGBToOKLCH

Introduced in: v25.7.0

Converts a colour encoded in the sRGB colour space to the perceptually uniform OKLCH colour space.

If any input channel is outside [0...255] or the gamma value is non-positive, the behaviour is implementation-defined.

OKLCH is a cylindrical version of the OKLab colour space. It’s three coordinates are L (the lightness in the range [0...1]), C (chroma >= 0) and H (the hue in degrees from [0...360]). OKLab/OKLCH is designed to be perceptually uniform while remaining cheap to compute.

The conversion consists of three stages:

1. sRGB to Linear sRGB
2. Linear sRGB to OKLab
3. OKLab to OKLCH.

For references of colors in the OKLCH space, and how they correspond to sRGB colors, please see https://OKLCH.com/.

Syntax

colorSRGBToOKLCH(tuple[, gamma])

Arguments

• tuple — Tuple of three values R, G, B in the range [0...255]. Tuple(UInt8, UInt8, UInt8)
• gamma — Optional. Exponent that is used to linearize sRGB by applying (x / 255)^gamma to each channel x. Defaults to 2.2. Float64

Returned value

Returns a tuple (L, C, H) representing the OKLCH color space values. Tuple(Float64, Float64, Float64)

Examples

Convert sRGB to OKLCH

Query

SELECT colorSRGBToOKLCH((128, 64, 32), 2.2) AS lch;

Response

┌─lch───────────────────────────────────────────────────────┐
│ (0.4436238384931984,0.1044269954567863,45.90734548193018) │
└───────────────────────────────────────────────────────────┘

connectionId

Introduced in: v21.3.0

Returns the connection ID of the client that submitted the current query. This function is most useful in debugging scenarios. It was created for compatibility with MySQL’s CONNECTION_ID function. It is not typically used in production queries.

Syntax

connectionId()

Arguments

• None.

Returned value

Returns the connection ID of the current client. UInt64

Examples

Usage example

Query

SELECT connectionId();

Response

┌─connectionId()─┐
│              0 │
└────────────────┘

countDigits

Introduced in: v20.8.0

Returns the number of decimal digits needed to represent a value.

This function takes into account the scales of decimal values i.e., it calculates the result over the underlying integer type which is (value * scale).

For example:

• countDigits(42) = 2
• countDigits(42.000) = 5
• countDigits(0.04200) = 4

You can check decimal overflow for Decimal64 with countDigits(x) > 18, although it is slower than isDecimalOverflow.

Syntax

countDigits(x)

Arguments

• x — An integer or decimal value. (U)Int* or Decimal

Returned value

Returns the number of digits needed to represent x. UInt8

Examples

Usage example

Query

SELECT countDigits(toDecimal32(1, 9)), countDigits(toDecimal32(-1, 9)),
       countDigits(toDecimal64(1, 18)), countDigits(toDecimal64(-1, 18)),
       countDigits(toDecimal128(1, 38)), countDigits(toDecimal128(-1, 38));

Response

┌─countDigits(toDecimal32(1, 9))─┬─countDigits(toDecimal32(-1, 9))─┬─countDigits(toDecimal64(1, 18))─┬─countDigits(toDecimal64(-1, 18))─┬─countDigits(toDecimal128(1, 38))─┬─countDigits(toDecimal128(-1, 38))─┐
│                             10 │                              10 │                              19 │                               19 │                               39 │                                39 │
└────────────────────────────────┴─────────────────────────────────┴─────────────────────────────────┴──────────────────────────────────┴──────────────────────────────────┴───────────────────────────────────┘

currentDatabase

Introduced in: v1.1.0

Returns the name of the current database. Useful in table engine parameters of CREATE TABLE queries where you need to specify the database.

Also see the SET statement.

Syntax

currentDatabase()

Aliases: current_database, SCHEMA, DATABASE

Arguments

• None.

Returned value

Returns the current database name. String

Examples

Usage example

Query

SELECT currentDatabase()

Response

┌─currentDatabase()─┐
│ default           │
└───────────────────┘

SQL standard syntax without parentheses

Query

SELECT CURRENT_DATABASE

Response

┌─CURRENT_DATABASE─┐
│ default          │
└──────────────────┘

currentProfiles

Introduced in: v21.9.0

Returns an array of the setting profiles for the current user.

Syntax

currentProfiles()

Arguments

• None.

Returned value

Returns an array of setting profiles for the current user. Array(String)

Examples

Usage example

Query

SELECT currentProfiles();

Response

┌─currentProfiles()─────────────────────────────┐
│ ['default', 'readonly_user', 'web_analytics'] │
└───────────────────────────────────────────────┘

currentQueryID

Introduced in: v25.2.0

Returns current Query id.

Syntax

currentQueryID()

Aliases: current_query_id

Arguments

• None.

Returned value

Examples

Example

Query

SELECT currentQueryID();

Response

┌─currentQueryID()─────────────────────┐
│ 1280d0e8-1a08-4524-be6e-77975bb68e7d │
└──────────────────────────────────────┘

currentRoles

Introduced in: v21.9.0

Returns an array of the roles which are assigned to the current user.

Syntax

currentRoles()

Arguments

• None.

Returned value

Returns an array of the roles which are assigned to the current user. Array(String)

Examples

Usage example

Query

SELECT currentRoles();

Response

┌─currentRoles()─────────────────────────────────┐
│ ['sql-console-role:[e-mail]'] │
└────────────────────────────────────────────────┘

currentSchemas

Introduced in: v23.7.0

Same as function currentDatabase but

• accepts a boolean argument which is ignored
• returns the database name as an array with a single value.

Function currentSchemas only exists for compatibility with PostgreSQL. Please use currentDatabase instead.

Also see the SET statement.

Syntax

currentSchemas(bool)

Aliases: current_schemas

Arguments

• bool — A boolean value, which is ignored. Bool

Returned value

Returns a single-element array with the name of the current database. Array(String)

Examples

Usage example

Query

SELECT currentSchemas(true)

Response

┌─currentSchemas(true)─┐
│ ['default']          │
└──────────────────────┘

currentUser

Introduced in: v20.1.0

Returns the name of the current user. In case of a distributed query, the name of the user who initiated the query is returned.

Syntax

currentUser()

Aliases: current_user, user

Arguments

• None.

Returned value

Returns the name of the current user, otherwise the login of the user who initiated the query. String

Examples

Usage example

Query

SELECT currentUser()

Response

┌─currentUser()─┐
│ default       │
└───────────────┘

SQL standard syntax without parentheses

Query

SELECT CURRENT_USER

Response

┌─CURRENT_USER─┐
│ default      │
└──────────────┘

defaultProfiles

Introduced in: v21.9.0

Returns an array of default setting profile names for the current user.

Syntax

defaultProfiles()

Arguments

• None.

Returned value

Returns an array of default setting profile names for the current user. Array(String)

Examples

Usage example

Query

SELECT defaultProfiles();

Response

┌─defaultProfiles()─┐
│ ['default']       │
└───────────────────┘

defaultRoles

Introduced in: v21.9.0

Returns an array of default roles for the current user.

Syntax

defaultRoles()

Arguments

• None.

Returned value

Returns an array of default roles for the current user. Array(String)

Examples

Usage example

Query

SELECT defaultRoles();

Response

┌─defaultRoles()─────────────────────────────────┐
│ ['sql-console-role:[e-mail]'] │
└────────────────────────────────────────────────┘

defaultValueOfArgumentType

Introduced in: v1.1.0

Returns the default value for a given data type. Does not include default values for custom columns set by the user.

Syntax

defaultValueOfArgumentType(expression)

Arguments

• expression — Arbitrary type of value or an expression that results in a value of an arbitrary type. Any

Returned value

Returns 0 for numbers, an empty string for strings or NULL for Nullable types. UInt8 or String or NULL

Examples

Usage example

Query

SELECT defaultValueOfArgumentType(CAST(1 AS Int8));

Response

┌─defaultValueOfArgumentType(CAST(1, 'Int8'))─┐
│                                           0 │
└─────────────────────────────────────────────┘

Nullable example

Query

SELECT defaultValueOfArgumentType(CAST(1 AS Nullable(Int8)));

Response

┌─defaultValueOfArgumentType(CAST(1, 'Nullable(Int8)'))─┐
│                                                  ᴺᵁᴸᴸ │
└───────────────────────────────────────────────────────┘

defaultValueOfTypeName

Introduced in: v1.1.0

Returns the default value for the given type name.

Syntax

defaultValueOfTypeName(type)

Arguments

• type — A string representing a type name. String

Returned value

Returns the default value for the given type name: 0 for numbers, an empty string for strings, or NULL for Nullable UInt8 or String or NULL

Examples

Usage example

Query

SELECT defaultValueOfTypeName('Int8');

Response

┌─defaultValueOfTypeName('Int8')─┐
│                              0 │
└────────────────────────────────┘

Nullable example

Query

SELECT defaultValueOfTypeName('Nullable(Int8)');

Response

┌─defaultValueOfTypeName('Nullable(Int8)')─┐
│                                     ᴺᵁᴸᴸ │
└──────────────────────────────────────────┘

displayName

Introduced in: v22.11.0

Returns the value of display_name from config or the server’s Fully Qualified Domain Name (FQDN) if not set.

Syntax

displayName()

Arguments

• None.

Returned value

Returns the value of display_name from config or server FQDN if not set. String

Examples

Usage example

Query

SELECT displayName();

Response

┌─displayName()─┐
│ production    │
└───────────────┘

dumpColumnStructure

Introduced in: v1.1.0

Outputs a detailed description of the internal structure of a column and its data type.

Syntax

dumpColumnStructure(x)

Arguments

• x — Value for which to get the description of. Any

Returned value

Returns a description of the column structure used for representing the value. String

Examples

Usage example

Query

SELECT dumpColumnStructure(CAST('2018-01-01 01:02:03', 'DateTime'));

Response

┌─dumpColumnStructure(CAST('2018-01-01 01:02:03', 'DateTime'))─┐
│ DateTime, Const(size = 1, UInt32(size = 1))                  │
└──────────────────────────────────────────────────────────────┘

enabledProfiles

Introduced in: v21.9.0

Returns an array of setting profile names which are enabled for the current user.

Syntax

enabledProfiles()

Arguments

• None.

Returned value

Returns an array of setting profile names which are enabled for the current user. Array(String)

Examples

Usage example

Query

SELECT enabledProfiles();

Response

┌─enabledProfiles()─────────────────────────────────────────────────┐
│ ['default', 'readonly_user', 'web_analytics', 'batch_processing'] │
└───────────────────────────────────────────────────────────────────┘

enabledRoles

Introduced in: v21.9.0

Returns an array of the roles which are enabled for the current user.

Syntax

enabledRoles()

Arguments

• None.

Returned value

Returns an array of role names which are enabled for the current user. Array(String)

Examples

Usage example

Query

SELECT enabledRoles();

Response

┌─enabledRoles()─────────────────────────────────────────────────┐
│ ['general_data', 'sql-console-role:[e-mail]'] │
└────────────────────────────────────────────────────────────────┘

errorCodeToName

Introduced in: v20.12.0

Returns the textual name of a numeric ClickHouse error code. The mapping from numeric error codes to error names is available here.

Syntax

errorCodeToName(error_code)

Arguments

• error_code — ClickHouse error code. (U)Int* or Float* or Decimal

Returned value

Returns the textual name of error_code. String

Examples

Usage example

Query

SELECT errorCodeToName(252);

Response

┌─errorCodeToName(252)─┐
│ TOO_MANY_PARTS       │
└──────────────────────┘

file

Introduced in: v21.3.0

Reads a file as a string and loads the data into the specified column. The file content is not interpreted.

Also see the file table function.

Syntax

file(path[, default])

Arguments

• path — The path of the file relative to the user_files_path. Supports wildcards *, **, ?, {abc,def} and {N..M} where N, M are numbers and 'abc', 'def' are strings. String
• default — The value returned if the file does not exist or cannot be accessed. String or NULL

Returned value

Returns the file content as a string. String

Examples

Insert files into a table

Query

INSERT INTO table SELECT file('a.txt'), file('b.txt');

Response

filesystemAvailable

Introduced in: v20.1.0

Returns the amount of free space in the filesystem hosting the database persistence. The returned value is always smaller than the total free space (filesystemUnreserved) because some space is reserved for the operating system.

Syntax

filesystemAvailable([disk_name])

Arguments

• disk_name — Optional. The disk name to find the amount of free space for. If omitted, uses the default disk. String or FixedString

Returned value

Returns the amount of remaining space available in bytes. UInt64

Examples

Usage example

Query

SELECT formatReadableSize(filesystemAvailable()) AS "Available space";

Response

┌─Available space─┐
│ 30.75 GiB       │
└─────────────────┘

filesystemCapacity

Introduced in: v20.1.0

Returns the capacity of the filesystem in bytes. Needs the path to the data directory to be configured.

Syntax

filesystemCapacity([disk_name])

Arguments

• disk_name — Optional. The disk name to get the capacity for. If omitted, uses the default disk. String or FixedString

Returned value

Returns the capacity of the filesystem in bytes. UInt64

Examples

Usage example

Query

SELECT formatReadableSize(filesystemCapacity()) AS "Capacity";

Response

┌─Capacity──┐
│ 39.32 GiB │
└───────────┘

filesystemUnreserved

Introduced in: v22.12.0

Returns the total amount of free space on the filesystem hosting the database persistence (previously filesystemFree). See also filesystemAvailable.

Syntax

filesystemUnreserved([disk_name])

Arguments

• disk_name — Optional. The disk name for which to find the total amount of free space. If omitted, uses the default disk. String or FixedString

Returned value

Returns the amount of free space in bytes. UInt64

Examples

Usage example

Query

SELECT formatReadableSize(filesystemUnreserved()) AS "Free space";

Response

┌─Free space─┐
│ 32.39 GiB  │
└────────────┘

finalizeAggregation

Introduced in: v1.1.0

Given an aggregation state, this function returns the result of aggregation (or the finalized state when using a -State combinator).

Syntax

finalizeAggregation(state)

Arguments

• state — State of aggregation. AggregateFunction

Returned value

Returns the finalized result of aggregation. Any

Examples

Usage example

Query

SELECT finalizeAggregation(arrayReduce('maxState', [1, 2, 3]));

Response

┌─finalizeAggregation(arrayReduce('maxState', [1, 2, 3]))─┐
│                                                       3 │
└─────────────────────────────────────────────────────────┘

Combined with initializeAggregation

Query

WITH initializeAggregation('sumState', number) AS one_row_sum_state
SELECT
    number,
    finalizeAggregation(one_row_sum_state) AS one_row_sum,
    runningAccumulate(one_row_sum_state) AS cumulative_sum
FROM numbers(5);

Response

┌─number─┬─one_row_sum─┬─cumulative_sum─┐
│      0 │           0 │              0 │
│      1 │           1 │              1 │
│      2 │           2 │              3 │
│      3 │           3 │              6 │
│      4 │           4 │             10 │
└────────┴─────────────┴────────────────┘

flipCoordinates

Introduced in: v25.11.0

Flips the x and y coordinates of geometric objects. This operation swaps latitude and longitude, which is useful for converting between different coordinate systems or correcting coordinate order.

For a Point, it swaps the x and y coordinates. For complex geometries (LineString, Polygon, MultiPolygon, Ring, MultiLineString), it recursively applies the transformation to each coordinate pair.

The function supports both individual geometry types (Point, Ring, Polygon, MultiPolygon, LineString, MultiLineString) and the Geometry variant type.

Syntax

flipCoordinates(geometry)

Arguments

• geometry — The geometry to transform. Supported types: Point (Tuple(Float64, Float64)), Ring (Array(Point)), Polygon (Array(Ring)), MultiPolygon (Array(Polygon)), LineString (Array(Point)), MultiLineString (Array(LineString)), or Geometry (a variant containing any of these types).

Returned value

The geometry with flipped coordinates. The return type matches the input type. Point or Ring or Polygon or MultiPolygon or LineString or MultiLineString or Geometry

Examples

basic_point

Query

SELECT flipCoordinates((1.0, 2.0));

Response

(2.0, 1.0)

ring

Query

SELECT flipCoordinates([(1.0, 2.0), (3.0, 4.0)]);

Response

[(2.0, 1.0), (4.0, 3.0)]

polygon

Query

SELECT flipCoordinates([[(1.0, 2.0), (3.0, 4.0)], [(5.0, 6.0), (7.0, 8.0)]]);

Response

[[(2.0, 1.0), (4.0, 3.0)], [(6.0, 5.0), (8.0, 7.0)]]

geometry_wkt

Query

SELECT flipCoordinates(readWkt('POINT(10 20)'));

Response

(20, 10)

geometry_polygon_wkt

Query

SELECT flipCoordinates(readWkt('POLYGON((0 0, 5 0, 5 5, 0 5, 0 0))'));

Response

[[(0, 0), (0, 5), (5, 5), (5, 0), (0, 0)]]

formatQuery

Introduced in: v23.10.0

Returns a formatted, possibly multi-line, version of the given SQL query. Throws in case of a parsing error. [example:multiline]

Syntax

formatQuery(query)

Arguments

• query — The SQL query to be formatted. String

Returned value

The formatted query String

Examples

multiline

Query

SELECT formatQuery('select a,    b FRom tab WHERE a > 3 and  b < 3');

Response

SELECT
    a,
    b
FROM tab
WHERE (a > 3) AND (b < 3)

formatQueryOrNull

Introduced in: v23.11.0

Returns a formatted, possibly multi-line, version of the given SQL query. Returns NULL in case of a parsing error. [example:multiline]

Syntax

formatQueryOrNull(query)

Arguments

• query — The SQL query to be formatted. String

Returned value

The formatted query String

Examples

multiline

Query

SELECT formatQuery('select a,    b FRom tab WHERE a > 3 and  b < 3');

Response

SELECT
    a,
    b
FROM tab
WHERE (a > 3) AND (b < 3)

formatQuerySingleLine

Introduced in: v23.10.0

Like formatQuery() but the returned formatted string contains no line breaks. Throws in case of a parsing error. [example:multiline]

Syntax

formatQuerySingleLine(query)

Arguments

• query — The SQL query to be formatted. String

Returned value

The formatted query String

Examples

multiline

Query

SELECT formatQuerySingleLine('select a,    b FRom tab WHERE a > 3 and  b < 3');

Response

SELECT a, b FROM tab WHERE (a > 3) AND (b < 3)

formatQuerySingleLineOrNull

Introduced in: v23.11.0

Like formatQuery() but the returned formatted string contains no line breaks. Returns NULL in case of a parsing error. [example:multiline]

Syntax

formatQuerySingleLineOrNull(query)

Arguments

• query — The SQL query to be formatted. String

Returned value

The formatted query String

Examples

multiline

Query

SELECT formatQuerySingleLine('select a,    b FRom tab WHERE a > 3 and  b < 3');

Response

SELECT a, b FROM tab WHERE (a > 3) AND (b < 3)

formatReadableDecimalSize

Introduced in: v22.11.0

Given a size (number of bytes), this function returns a readable, rounded size with suffix (KB, MB, etc.) as a string.

The opposite operations of this function are parseReadableSize.

Syntax

formatReadableDecimalSize(x)

Arguments

• x — Size in bytes. UInt64

Returned value

Returns a readable, rounded size with suffix as a string. String

Examples

Format file sizes

Query

SELECT
    arrayJoin([1, 1024, 1024*1024, 192851925]) AS filesize_bytes,
    formatReadableDecimalSize(filesize_bytes) AS filesize

Response

┌─filesize_bytes─┬─filesize───┐
│              1 │ 1.00 B     │
│           1024 │ 1.02 KB    │
│        1048576 │ 1.05 MB    │
│      192851925 │ 192.85 MB  │
└────────────────┴────────────┘

formatReadableQuantity

Introduced in: v20.10.0

Given a number, this function returns a rounded number with suffix (thousand, million, billion, etc.) as a string.

This function accepts any numeric type as input, but internally it casts them to Float64. Results might be suboptimal with large values.

Syntax

formatReadableQuantity(x)

Arguments

• x — A number to format. UInt64

Returned value

Returns a rounded number with suffix as a string. String

Examples

Format numbers with suffixes

Query

SELECT
    arrayJoin([1024, 1234 * 1000, (4567 * 1000) * 1000, 98765432101234]) AS number,
    formatReadableQuantity(number) AS number_for_humans

Response

┌─────────number─┬─number_for_humans─┐
│           1024 │ 1.02 thousand     │
│        1234000 │ 1.23 million      │
│     4567000000 │ 4.57 billion      │
│ 98765432101234 │ 98.77 trillion    │
└────────────────┴───────────────────┘

formatReadableSize

Introduced in: v1.1.0

Given a size (number of bytes), this function returns a readable, rounded size with suffix (KiB, MiB, etc.) as string.

The opposite operations of this function are parseReadableSize, parseReadableSizeOrZero, and parseReadableSizeOrNull. This function accepts any numeric type as input, but internally it casts them to Float64. Results might be suboptimal with large values.

Syntax

formatReadableSize(x)

Aliases: FORMAT_BYTES

Arguments

• x — Size in bytes. UInt64

Returned value

Returns a readable, rounded size with suffix as a string. String

Examples

Format file sizes

Query

SELECT
    arrayJoin([1, 1024, 1024*1024, 192851925]) AS filesize_bytes,
    formatReadableSize(filesize_bytes) AS filesize

Response

┌─filesize_bytes─┬─filesize───┐
│              1 │ 1.00 B     │
│           1024 │ 1.00 KiB   │
│        1048576 │ 1.00 MiB   │
│      192851925 │ 183.92 MiB │
└────────────────┴────────────┘

formatReadableTimeDelta

Introduced in: v20.12.0

Given a time interval (delta) in seconds, this function returns a time delta with year/month/day/hour/minute/second/millisecond/microsecond/nanosecond as a string.

This function accepts any numeric type as input, but internally it casts them to Float64. Results might be suboptimal with large values.

Syntax

formatReadableTimeDelta(column[, maximum_unit, minimum_unit])

Arguments

• column — A column with a numeric time delta. Float64
• maximum_unit — Optional. Maximum unit to show. Acceptable values: nanoseconds, microseconds, milliseconds, seconds, minutes, hours, days, months, years. Default value: years. const String
• minimum_unit — Optional. Minimum unit to show. All smaller units are truncated. Acceptable values: nanoseconds, microseconds, milliseconds, seconds, minutes, hours, days, months, years. If explicitly specified value is bigger than maximum_unit, an exception will be thrown. Default value: seconds if maximum_unit is seconds or bigger, nanoseconds otherwise. const String

Returned value

Returns a time delta as a string. String

Examples

Usage example

Query

SELECT
    arrayJoin([100, 12345, 432546534]) AS elapsed,
    formatReadableTimeDelta(elapsed) AS time_delta

Response

┌────elapsed─┬─time_delta─────────────────────────────────────────────────────┐
│        100 │ 1 minute and 40 seconds                                        │
│      12345 │ 3 hours, 25 minutes and 45 seconds                             │
│  432546534 │ 13 years, 8 months, 17 days, 7 hours, 48 minutes and 54 seconds│
└────────────┴────────────────────────────────────────────────────────────────┘

With maximum unit

Query

SELECT
    arrayJoin([100, 12345, 432546534]) AS elapsed,
    formatReadableTimeDelta(elapsed, 'minutes') AS time_delta

Response

┌────elapsed─┬─time_delta─────────────────────────────────────────────────────┐
│        100 │ 1 minute and 40 seconds                                         │
│      12345 │ 205 minutes and 45 seconds                                      │
│  432546534 │ 7209108 minutes and 54 seconds                                  │
└────────────┴─────────────────────────────────────────────────────────────────┘

fuzzQuery

Introduced in: v26.2.0

Parses the given query string and applies random AST mutations (fuzzing) to it. Returns the fuzzed query as a string. Non-deterministic: each call may produce a different result. Requires allow_fuzz_query_functions = 1.

Syntax

fuzzQuery(query)

Arguments

• query — The SQL query to be fuzzed. String

Returned value

The fuzzed query string String

Examples

basic

Query

SET allow_fuzz_query_functions = 1; SELECT fuzzQuery('SELECT 1');

Response

generateRandomStructure

Introduced in: v23.5.0

Generates random table structure in the format column1_name column1_type, column2_name column2_type, ....

Syntax

generateRandomStructure([number_of_columns, seed])

Arguments

• number_of_columns — The desired number of columns in the resultant table structure. If set to 0 or Null, the number of columns will be random from 1 to 128. Default value: Null. UInt64
• seed — Random seed to produce stable results. If seed is not specified or set to Null, it is randomly generated. UInt64

Returned value

Randomly generated table structure. String

Examples

Usage example

Query

SELECT generateRandomStructure()

Response

c1 Decimal32(5), c2 Date, c3 Tuple(LowCardinality(String), Int128, UInt64, UInt16, UInt8, IPv6), c4 Array(UInt128), c5 UInt32, c6 IPv4, c7 Decimal256(64), c8 Decimal128(3), c9 UInt256, c10 UInt64, c11 DateTime

with specified number of columns

Query

SELECT generateRandomStructure(1)

Response

c1 Map(UInt256, UInt16)

with specified seed

Query

SELECT generateRandomStructure(NULL, 33)

Response

c1 DateTime, c2 Enum8('c2V0' = 0, 'c2V1' = 1, 'c2V2' = 2, 'c2V3' = 3), c3 LowCardinality(Nullable(FixedString(30))), c4 Int16, c5 Enum8('c5V0' = 0, 'c5V1' = 1, 'c5V2' = 2, 'c5V3' = 3), c6 Nullable(UInt8), c7 String, c8 Nested(e1 IPv4, e2 UInt8, e3 UInt16, e4 UInt16, e5 Int32, e6 Map(Date, Decimal256(70)))

generateSerialID

Introduced in: v25.1.0

Generates and returns sequential numbers starting from the previous counter value. This function takes a string argument - a series identifier, and an optional starting value. The server should be configured with Keeper. The series are stored in Keeper nodes under the path, which can be configured in series_keeper_path in the server configuration.

Syntax

generateSerialID(series_identifier[, start_value])

Arguments

• series_identifier — Series identifier const String
• start_value — Optional. Starting value for the counter. Defaults to 0. Note: this value is only used when creating a new series and is ignored if the series already exists UInt*

Returned value

Returns sequential numbers starting from the previous counter value. UInt64

Examples

first call

Query

SELECT generateSerialID('id1')

Response

┌─generateSerialID('id1')──┐
│                        1 │
└──────────────────────────┘

second call

Query

SELECT generateSerialID('id1')

Response

┌─generateSerialID('id1')──┐
│                        2 │
└──────────────────────────┘

column call

Query

SELECT *, generateSerialID('id1') FROM test_table

Response

┌─CounterID─┬─UserID─┬─ver─┬─generateSerialID('id1')──┐
│         1 │      3 │   3 │                        3 │
│         1 │      1 │   1 │                        4 │
│         1 │      2 │   2 │                        5 │
│         1 │      5 │   5 │                        6 │
│         1 │      4 │   4 │                        7 │
└───────────┴────────┴─────┴──────────────────────────┘

with start value

Query

SELECT generateSerialID('id2', 100)

Response

┌─generateSerialID('id2', 100)──┐
│                           100 │
└───────────────────────────────┘

with start value second call

Query

SELECT generateSerialID('id2', 100)

Response

┌─generateSerialID('id2', 100)──┐
│                           101 │
└───────────────────────────────┘

getClientHTTPHeader

Introduced in: v24.5.0

Gets the value of an HTTP header. If there is no such header or the current request is not performed via the HTTP interface, the function returns an empty string. Certain HTTP headers (e.g., Authentication and X-ClickHouse-*) are restricted.

Setting allow_get_client_http_header is required

The function requires the setting allow_get_client_http_header to be enabled. The setting is not enabled by default for security reasons, because some headers, such as Cookie, could contain sensitive info.

HTTP headers are case sensitive for this function. If the function is used in the context of a distributed query, it returns non-empty result only on the initiator node.

Syntax

getClientHTTPHeader(name)

Arguments

• name — The HTTP header name. String

Returned value

Returns the value of the header. String

Examples

Usage example

Query

SELECT getClientHTTPHeader('Content-Type');

Response

┌─getClientHTTPHeader('Content-Type')─┐
│ application/x-www-form-urlencoded   │
└─────────────────────────────────────┘

getMacro

Introduced in: v20.1.0

Returns the value of a macro from the server configuration file. Macros are defined in the <macros> section of the configuration file and can be used to distinguish servers by convenient names even if they have complicated hostnames. If the function is executed in the context of a distributed table, it generates a normal column with values relevant to each shard.

Syntax

getMacro(name)

Arguments

• name — The name of the macro to retrieve. const String

Returned value

Returns the value of the specified macro. String

Examples

Basic usage

Query

SELECT getMacro('test');

Response

┌─getMacro('test')─┐
│ Value            │
└──────────────────┘

getMaxTableNameLengthForDatabase

Introduced in: v25.1.0

Returns the maximum table name length in a specified database.

Syntax

getMaxTableNameLengthForDatabase(database_name)

Arguments

• database_name — The name of the specified database. String

Returned value

Returns the length of the maximum table name, an Integer

Examples

typical

Query

SELECT getMaxTableNameLengthForDatabase('default');

Response

┌─getMaxTableNameLengthForDatabase('default')─┐
            │                                         206 │
            └─────────────────────────────────────────────┘

getMergeTreeSetting

Introduced in: v25.6.0

Returns the current value of a MergeTree setting.

Syntax

getMergeTreeSetting(setting_name)

Arguments

• setting_name — The setting name. String

Returned value

Returns the merge tree setting’s current value.

Examples

Usage example

Query

SELECT getMergeTreeSetting('index_granularity');

Response

┌─getMergeTreeSetting('index_granularity')─┐
│                                     8192 │
└──────────────────────────────────────────┘

getOSKernelVersion

Introduced in: v21.11.0

Returns a string with the OS kernel version.

Syntax

getOSKernelVersion()

Arguments

• None.

Returned value

Returns the current OS kernel version. String

Examples

Usage example

Query

SELECT getOSKernelVersion();

Response

┌─getOSKernelVersion()────┐
│ Linux 4.15.0-55-generic │
└─────────────────────────┘

getServerPort

Introduced in: v21.10.0

Returns the server’s port number for a given protocol.

Syntax

getServerPort(port_name)

Arguments

• port_name — The name of the port. String

Returned value

Returns the server port number. UInt16

Examples

Usage example

Query

SELECT getServerPort('tcp_port');

Response

┌─getServerPort('tcp_port')─┐
│                      9000 │
└───────────────────────────┘

getServerSetting

Introduced in: v25.6.0

Returns the currently set value, given a server setting name.

Syntax

getServerSetting(setting_name')

Arguments

• setting_name — The server setting name. String

Returned value

Returns the server setting’s current value. Any

Examples

Usage example

Query

SELECT getServerSetting('allow_use_jemalloc_memory');

Response

┌─getServerSetting('allow_use_jemalloc_memory')─┐
│ true                                          │
└───────────────────────────────────────────────┘

getSetting

Introduced in: v20.7.0

Returns the current value of a setting.

Syntax

getSetting(setting_name)

Arguments

• setting_Name — The setting name. const String

Returned value

Returns the setting’s current value. Any

Examples

Usage example

Query

SELECT getSetting('enable_analyzer');
SET enable_analyzer = false;
SELECT getSetting('enable_analyzer');

Response

┌─getSetting('⋯_analyzer')─┐
│ true                     │
└──────────────────────────┘
┌─getSetting('⋯_analyzer')─┐
│ false                    │
└──────────────────────────┘

getSettingOrDefault

Introduced in: v24.10.0

Returns the current value of a setting or returns the default value specified in the second argument if the setting is not set in the current profile.

Syntax

getSettingOrDefault(setting_name, default_value)

Arguments

• setting_name — The setting name. String
• default_value — Value to return if custom_setting is not set. Value may be of any data type or Null.

Returned value

Returns the current value of the specified setting or default_value if the setting is not set.

Examples

Usage example

Query

SELECT getSettingOrDefault('custom_undef1', 'my_value');
SELECT getSettingOrDefault('custom_undef2', 100);
SELECT getSettingOrDefault('custom_undef3', NULL);

Response

my_value
100
NULL

getSizeOfEnumType

Introduced in: v1.1.0

Returns the number of fields in the given Enum.

Syntax

getSizeOfEnumType(x)

Arguments

• x — Value of type Enum. Enum

Returned value

Returns the number of fields with Enum input values. UInt8/16

Examples

Usage example

Query

SELECT getSizeOfEnumType(CAST('a' AS Enum8('a' = 1, 'b' = 2))) AS x;

Response

┌─x─┐
│ 2 │
└───┘

getSubcolumn

Introduced in: v23.3.0

Receives the expression or identifier and constant string with the name of subcolumn.

Returns requested subcolumn extracted from the expression.

Syntax

getSubcolumn(nested_value, subcolumn_name)

Arguments

• None.

Returned value

Examples

getSubcolumn

Query

SELECT getSubcolumn(array_col, 'size0'), getSubcolumn(tuple_col, 'elem_name')

Response

getTypeSerializationStreams

Introduced in: v22.6.0

Enumerates stream paths of a data type. This function is intended for developmental use.

Syntax

getTypeSerializationStreams(col)

Arguments

• col — Column or string representation of a data-type from which the data type will be detected. Any

Returned value

Returns an array with all the serialization sub-stream paths. Array(String)

Examples

tuple

Query

SELECT getTypeSerializationStreams(tuple('a', 1, 'b', 2))

Response

['{TupleElement(1), Regular}','{TupleElement(2), Regular}','{TupleElement(3), Regular}','{TupleElement(4), Regular}']

map

Query

SELECT getTypeSerializationStreams('Map(String, Int64)')

Response

['{ArraySizes}','{ArrayElements, TupleElement(keys), Regular}','{ArrayElements, TupleElement(values), Regular}']

globalVariable

Introduced in: v20.5.0

Takes a constant string argument and returns the value of the global variable with that name. This function is intended for compatibility with MySQL and not needed or useful for normal operation of ClickHouse. Only few dummy global variables are defined.

Syntax

globalVariable(name)

Arguments

• name — Global variable name. String

Returned value

Returns the value of variable name. Any

Examples

globalVariable

Query

SELECT globalVariable('max_allowed_packet')

Response

67108864

hasColumnInTable

Introduced in: v1.1.0

Checks if a specific column exists in a database table. For elements in a nested data structure, the function checks for the existence of a column. For the nested data structure itself, the function returns 0.

Syntax

hasColumnInTable([hostname[, username[, password]],]database, table, column)

Arguments

• database — Name of the database. const String
• table — Name of the table. const String
• column — Name of the column. const String
• hostname — Optional. Remote server name to perform the check on. const String
• username — Optional. Username for remote server. const String
• password — Optional. Password for remote server. const String

Returned value

Returns 1 if the given column exists, 0 otherwise. UInt8

Examples

Check an existing column

Query

SELECT hasColumnInTable('system','metrics','metric')

Response

1

Check a non-existing column

Query

SELECT hasColumnInTable('system','metrics','non-existing_column')

Response

0

hasThreadFuzzer

Introduced in: v20.6.0

Returns whether the thread fuzzer is enabled. THis function is only useful for testing and debugging.

Syntax

hasThreadFuzzer()

Arguments

• None.

Returned value

Returns whether Thread Fuzzer is effective. UInt8

Examples

Check Thread Fuzzer status

Query

SELECT hasThreadFuzzer()

Response

┌─hasThreadFuzzer()─┐
│                 0 │
└───────────────────┘

highlightQuery

Introduced in: v26.5.0

Parses a ClickHouse SQL query string and returns an array of highlighted ranges for syntax highlighting. Each range is a named tuple with the beginning position (in bytes), the end position, and the highlight type. The highlight types describe the syntactic role of the fragment (keyword, identifier, function, etc.) and can be used to assign colors in a UI. Inside LIKE and REGEXP string patterns, metacharacters and escape characters are highlighted separately.

Syntax

highlightQuery(query)

Arguments

• query — A ClickHouse SQL query string. String.

Returned value

An array of named tuples (begin UInt64, end UInt64, type Enum8(...)) representing highlighted ranges. Array(Tuple(begin UInt64, end UInt64, type Enum8(...)))

Examples

simple

Query

SELECT highlightQuery('SELECT 1')

Response

[(0,6,'keyword'),(7,8,'number')]

hostName

Introduced in: v20.5.0

Returns the name of the host on which this function was executed. If the function executes on a remote server (distributed processing), the remote server name is returned. If the function executes in the context of a distributed table, it generates a normal column with values relevant to each shard. Otherwise it produces a constant value.

Syntax

hostName()

Aliases: hostname

Arguments

• None.

Returned value

Returns the host name. String

Examples

Usage example

Query

SELECT hostName()

Response

┌─hostName()─┐
│ clickhouse │
└────────────┘

icebergBucket

Introduced in: v25.5.0

Implements logic for the iceberg bucket transform

Syntax

icebergBucket(N, value)

Arguments

• N — The number of buckets, modulo. const (U)Int*
• value — The source value to transform. (U)Int* or Bool or Decimal or Float* or String or FixedString or UUID or Date or Time or DateTime

Returned value

Returns a 32-bit hash of the source value. Int32

Examples

Example

Query

SELECT icebergBucket(5, 1.0 :: Float32)

Response

4

icebergTruncate

Introduced in: v25.3.0

Implements logic of iceberg truncate transform: https://iceberg.apache.org/spec/#truncate-transform-details.

Syntax

icebergTruncate(N, value)

Arguments

• value — The value to transform. String or (U)Int* or Decimal

Returned value

The same type as the argument

Examples

Example

Query

SELECT icebergTruncate(3, 'iceberg')

Response

ice

identity

Introduced in: v1.1.0

This function returns the argument you pass to it, which is useful for debugging and testing. It lets you bypass index usage to see full scan performance instead. The query analyzer ignores anything inside identity functions when looking for indexes to use, and it also disables constant folding.

Syntax

identity(x)

Arguments

• x — Input value. Any

Returned value

Returns the input value unchanged. Any

Examples

Usage example

Query

SELECT identity(42)

Response

42

ignore

Introduced in: v1.1.0

Accepts arbitrary arguments and unconditionally returns 0.

Syntax

ignore(x)

Arguments

• x — An input value which is unused and passed only so as to avoid a syntax error. Any

Returned value

Always returns 0. UInt8

Examples

Usage example

Query

SELECT ignore(0, 'ClickHouse', NULL)

Response

┌─ignore(0, 'ClickHouse', NULL)─┐
│                             0 │
└───────────────────────────────┘

indexHint

Introduced in: v1.1.0

This function is intended for debugging and introspection. It ignores its argument and always returns 1. The arguments are not evaluated.

During index analysis, the argument of this function is assumed to not be wrapped in indexHint. This allows you to select data in index ranges by the corresponding condition but without further filtering by this condition. The index in ClickHouse is sparse and using indexHint will yield more data than specifying the same condition directly.

Explanation

When you run:

SELECT * FROM test WHERE key = 123;

ClickHouse does two things:

1. Uses the index to find which granules (blocks of ~8192 rows) might contain key = 123
2. Reads those granules and filters them row-by-row to return only rows where key = 123

So even if it reads 8,192 rows from disk, it only returns the 1 row that actually matches.

With indexHint, when you run:

SELECT * FROM test WHERE indexHint(key = 123);

ClickHouse does only one thing:

1. Uses the index to find which granules might contain key = 123 and returns all rows from those granules without filtering.

It returns all 8,192 rows, including rows where key = 456, key = 789, etc. (Everything that happened to be stored in the same granule.) indexHint() is not for performance. It’s for debugging and understanding how ClickHouse’s index works:

• Which granules does my condition select?
• How many rows are in those granules?
• Is my index being used effectively?

Note: It is not possible to optimize a query with the indexHint function. The indexHint function does not optimize the query, as it does not provide any additional information for the query analysis. Having an expression inside the indexHint function is not anyhow better than without the indexHint function. The indexHint function can be used only for introspection and debugging purposes and it does not improve performance. If you see the usage of indexHint by anyone other than ClickHouse contributors, it is likely a mistake and you should remove it.

Syntax

indexHint(expression)

Arguments

• expression — Any expression for index range selection. Expression

Returned value

Returns 1 in all cases. UInt8

Examples

Usage example with date filtering

Query

SELECT FlightDate AS k, count() FROM ontime WHERE indexHint(k = '2025-09-15') GROUP BY k ORDER BY k ASC;

Response

┌──────────k─┬─count()─┐
│ 2025-09-14 │    7071 │
│ 2025-09-15 │   16428 │
│ 2025-09-16 │    1077 │
│ 2025-09-30 │    8167 │
└────────────┴─────────┘

initialQueryID

Introduced in: v1.1.0

Returns the ID of the initial current query. Other parameters of a query can be extracted from field initial_query_id in system.query_log.

In contrast to queryID function, initialQueryID returns the same results on different shards.

Syntax

initialQueryID()

Aliases: initial_query_id

Arguments

• None.

Returned value

Returns the ID of the initial current query. String

Examples

Usage example

Query

CREATE TABLE tmp (str String) ENGINE = Log;
INSERT INTO tmp (*) VALUES ('a');
SELECT count(DISTINCT t) FROM (SELECT initialQueryID() AS t FROM remote('127.0.0.{1..3}', currentDatabase(), 'tmp') GROUP BY queryID());

Response

┌─count(DISTINCT t)─┐
│                 1 │
└───────────────────┘

initialQueryStartTime

Introduced in: v25.4.0

Returns the start time of the initial current query. initialQueryStartTime returns the same results on different shards.

Syntax

initialQueryStartTime()

Aliases: initial_query_start_time

Arguments

• None.

Returned value

Returns the start time of the initial current query. DateTime

Examples

Usage example

Query

CREATE TABLE tmp (str String) ENGINE = Log;
INSERT INTO tmp (*) VALUES ('a');
SELECT count(DISTINCT t) FROM (SELECT initialQueryStartTime() AS t FROM remote('127.0.0.{1..3}', currentDatabase(), 'tmp') GROUP BY queryID());

Response

┌─count(DISTINCT t)─┐
│                 1 │
└───────────────────┘

initializeAggregation

Introduced in: v20.6.0

Calculates the result of an aggregate function based on a single value. This function can be used to initialize aggregate functions with combinator -State. You can create states of aggregate functions and insert them to columns of type AggregateFunction or use initialized aggregates as default values.

Syntax

initializeAggregation(aggregate_function, arg1[, arg2, ...])

Arguments

• aggregate_function — Name of the aggregation function to initialize. String
• arg1[, arg2, ...] — Arguments of the aggregate function. Any

Returned value

Returns the result of aggregation for every row passed to the function. The return type is the same as the return type of the function that initializeAggregation takes as a first argument. Any

Examples

Basic usage with uniqState

Query

SELECT uniqMerge(state) FROM (SELECT initializeAggregation('uniqState', number % 3) AS state FROM numbers(10000));

Response

┌─uniqMerge(state)─┐
│                3 │
└──────────────────┘

Usage with sumState and finalizeAggregation

Query

SELECT finalizeAggregation(state), toTypeName(state) FROM (SELECT initializeAggregation('sumState', number % 3) AS state FROM numbers(5));

Response

┌─finalizeAggregation(state)─┬─toTypeName(state)─────────────┐
│                          0 │ AggregateFunction(sum, UInt8) │
│                          1 │ AggregateFunction(sum, UInt8) │
│                          2 │ AggregateFunction(sum, UInt8) │
│                          0 │ AggregateFunction(sum, UInt8) │
│                          1 │ AggregateFunction(sum, UInt8) │
└────────────────────────────┴───────────────────────────────┘

isConstant

Introduced in: v20.3.0

Returns whether the argument is a constant expression. A constant expression is an expression whose result is known during query analysis, i.e. before execution. For example, expressions over literals are constant expressions. This function is mostly intended for development, debugging and demonstration.

Syntax

isConstant(x)

Arguments

• x — An expression to check. Any

Returned value

Returns 1 if x is constant, 0 if x is non-constant. UInt8

Examples

Constant expression

Query

SELECT isConstant(x + 1)
FROM (SELECT 43 AS x)

Response

┌─isConstant(plus(x, 1))─┐
│                      1 │
└────────────────────────┘

Constant with function

Query

WITH 3.14 AS pi
SELECT isConstant(cos(pi))

Response

┌─isConstant(cos(pi))─┐
│                   1 │
└─────────────────────┘

Non-constant expression

Query

SELECT isConstant(number)
FROM numbers(1)

Response

┌─isConstant(number)─┐
│                  0 │
└────────────────────┘

Behavior of the now() function

Query

SELECT isConstant(now())

Response

┌─isConstant(now())─┐
│                 1 │
└───────────────────┘

isDecimalOverflow

Introduced in: v20.8.0

Checks if a decimal number has too many digits to fit properly in a Decimal data type with given precision.

Syntax

isDecimalOverflow(value[, precision])

Arguments

• value — Decimal value to check. Decimal
• precision — Optional. The precision of the Decimal type. If omitted, the initial precision of the first argument is used. UInt8

Returned value

Returns 1 if the decimal value has more digits than allowed by its precision, 0 if the decimal value satisfies the specified precision. UInt8

Examples

Usage example

Query

SELECT isDecimalOverflow(toDecimal32(1000000000, 0), 9),
       isDecimalOverflow(toDecimal32(1000000000, 0)),
       isDecimalOverflow(toDecimal32(-1000000000, 0), 9),
       isDecimalOverflow(toDecimal32(-1000000000, 0));

Response

┌─isDecimalOverflow(toDecimal32(1000000000, 0), 9)─┬─isDecimalOverflow(toDecimal32(1000000000, 0))─┬─isDecimalOverflow(toDecimal32(-1000000000, 0), 9)─┬─isDecimalOverflow(toDecimal32(-1000000000, 0))─┐
│                                                1 │                                             1 │                                                 1 │                                              1 │
└──────────────────────────────────────────────────┴───────────────────────────────────────────────┴───────────────────────────────────────────────────┴────────────────────────────────────────────────┘

joinGet

Introduced in: v18.16.0

Allows you to extract data from a table the same way as from a dictionary. Gets data from Join tables using the specified join key.

Only supports tables created with the ENGINE = Join(ANY, LEFT, <join_keys>) statement.

Syntax

joinGet(join_storage_table_name, value_column, join_keys)

Arguments

• join_storage_table_name — An identifier which indicates where to perform the search. The identifier is searched in the default database (see parameter default_database in the config file). To override the default database, use the USE database_name query or specify the database and the table through a dot, like database_name.table_name. String
• value_column — The name of the column of the table that contains required data. const String
• join_keys — A list of join keys. Any

Returned value

Returns list of values corresponded to list of keys. Any

Examples

Usage example

Query

CREATE TABLE db_test.id_val(`id` UInt32, `val` UInt32) ENGINE = Join(ANY, LEFT, id);
INSERT INTO db_test.id_val VALUES (1,11)(2,12)(4,13);

SELECT joinGet(db_test.id_val, 'val', toUInt32(1));

Response

┌─joinGet(db_test.id_val, 'val', toUInt32(1))─┐
│                                          11 │
└─────────────────────────────────────────────┘

Usage with table from current database

Query

USE db_test;
SELECT joinGet(id_val, 'val', toUInt32(2));

Response

┌─joinGet(id_val, 'val', toUInt32(2))─┐
│                                  12 │
└─────────────────────────────────────┘

Using arrays as join keys

Query

CREATE TABLE some_table (id1 UInt32, id2 UInt32, name String) ENGINE = Join(ANY, LEFT, id1, id2);
INSERT INTO some_table VALUES (1, 11, 'a') (2, 12, 'b') (3, 13, 'c');

SELECT joinGet(some_table, 'name', 1, 11);

Response

┌─joinGet(some_table, 'name', 1, 11)─┐
│ a                                  │
└────────────────────────────────────┘

joinGetOrNull

Introduced in: v20.4.0

Allows you to extract data from a table the same way as from a dictionary. Gets data from Join tables using the specified join key. Unlike joinGet it returns NULL when the key is missing.

Only supports tables created with the ENGINE = Join(ANY, LEFT, <join_keys>) statement.

Syntax

joinGetOrNull(join_storage_table_name, value_column, join_keys)

Arguments

• join_storage_table_name — An identifier which indicates where to perform the search. The identifier is searched in the default database (see parameter default_database in the config file). To override the default database, use the USE database_name query or specify the database and the table through a dot, like database_name.table_name. String
• value_column — The name of the column of the table that contains required data. const String
• join_keys — A list of join keys. Any

Returned value

Returns a list of values corresponding to the list of keys, or NULL if a key is not found. Any

Examples

Usage example

Query

CREATE TABLE db_test.id_val(`id` UInt32, `val` UInt32) ENGINE = Join(ANY, LEFT, id);
INSERT INTO db_test.id_val VALUES (1,11)(2,12)(4,13);

SELECT joinGetOrNull(db_test.id_val, 'val', toUInt32(1)), joinGetOrNull(db_test.id_val, 'val', toUInt32(999));

Response

┌─joinGetOrNull(db_test.id_val, 'val', toUInt32(1))─┬─joinGetOrNull(db_test.id_val, 'val', toUInt32(999))─┐
│                                                11 │                                                ᴺᵁᴸᴸ │
└───────────────────────────────────────────────────┴─────────────────────────────────────────────────────┘

lowCardinalityIndices

Introduced in: v18.12.0

Returns the position of a value in the dictionary of a LowCardinality column. Positions start at 1. Since LowCardinality have per-part dictionaries, this function may return different positions for the same value in different parts.

Syntax

lowCardinalityIndices(col)

Arguments

• col — A low cardinality column. LowCardinality

Returned value

The position of the value in the dictionary of the current part. UInt64

Examples

Usage examples

Query

DROP TABLE IF EXISTS test;
CREATE TABLE test (s LowCardinality(String)) ENGINE = Memory;

-- create two parts:

INSERT INTO test VALUES ('ab'), ('cd'), ('ab'), ('ab'), ('df');
INSERT INTO test VALUES ('ef'), ('cd'), ('ab'), ('cd'), ('ef');

SELECT s, lowCardinalityIndices(s) FROM test;

Response

┌─s──┬─lowCardinalityIndices(s)─┐
│ ab │                        1 │
│ cd │                        2 │
│ ab │                        1 │
│ ab │                        1 │
│ df │                        3 │
└────┴──────────────────────────┘
┌─s──┬─lowCardinalityIndices(s)─┐
│ ef │                        1 │
│ cd │                        2 │
│ ab │                        3 │
│ cd │                        2 │
│ ef │                        1 │
└────┴──────────────────────────┘

lowCardinalityKeys

Introduced in: v18.12.0

Returns the dictionary values of a LowCardinality column. If the block is smaller or larger than the dictionary size, the result will be truncated or extended with default values. Since LowCardinality have per-part dictionaries, this function may return different dictionary values in different parts.

Syntax

lowCardinalityKeys(col)

Arguments

• col — A low cardinality column. LowCardinality

Returned value

Returns the dictionary keys. UInt64

Examples

lowCardinalityKeys

Query

DROP TABLE IF EXISTS test;
CREATE TABLE test (s LowCardinality(String)) ENGINE = Memory;

-- create two parts:

INSERT INTO test VALUES ('ab'), ('cd'), ('ab'), ('ab'), ('df');
INSERT INTO test VALUES ('ef'), ('cd'), ('ab'), ('cd'), ('ef');

SELECT s, lowCardinalityKeys(s) FROM test;

Response

┌─s──┬─lowCardinalityKeys(s)─┐
│ ef │                       │
│ cd │ ef                    │
│ ab │ cd                    │
│ cd │ ab                    │
│ ef │                       │
└────┴───────────────────────┘
┌─s──┬─lowCardinalityKeys(s)─┐
│ ab │                       │
│ cd │ ab                    │
│ ab │ cd                    │
│ ab │ df                    │
│ df │                       │
└────┴───────────────────────┘

materialize

Introduced in: v1.1.0

Turns a constant into a full column containing a single value. Full columns and constants are represented differently in memory. Functions usually execute different code for normal and constant arguments, although the result should typically be the same. This function can be used to debug this behavior.

Syntax

materialize(x)

Arguments

• x — A constant. Any

Returned value

Returns a full column containing the constant value. Any

Examples

Usage example

Query

-- In the example below the `countMatches` function expects a constant second argument.
-- This behaviour can be debugged by using the `materialize` function to turn a constant into a full column,
-- verifying that the function throws an error for a non-constant argument.

SELECT countMatches('foobarfoo', 'foo');
SELECT countMatches('foobarfoo', materialize('foo'));

Response

2
Code: 44. DB::Exception: Received from localhost:9000. DB::Exception: Illegal type of argument #2 'pattern' of function countMatches, expected constant String, got String

minSampleSizeContinuous

Introduced in: v23.10.0

Calculates the minimum required sample size for an A/B test comparing means of a continuous metric in two samples.

Uses the formula described in this article. Assumes equal sizes of treatment and control groups. Returns the required sample size for one group (i.e. the sample size required for the whole experiment is twice the returned value). Also assumes equal variance of the test metric in treatment and control groups.

Syntax

minSampleSizeContinuous(baseline, sigma, mde, power, alpha)

Aliases: minSampleSizeContinous

Arguments

• baseline — Baseline value of a metric. (U)Int* or Float*
• sigma — Baseline standard deviation of a metric. (U)Int* or Float*
• mde — Minimum detectable effect (MDE) as percentage of the baseline value (e.g. for a baseline value 112.25 the MDE 0.03 means an expected change to 112.25 ± 112.25*0.03). (U)Int* or Float*
• power — Required statistical power of a test (1 - probability of Type II error). (U)Int* or Float*
• alpha — Required significance level of a test (probability of Type I error). (U)Int* or Float*

Returned value

Returns a named Tuple with 3 elements: minimum_sample_size, detect_range_lower and detect_range_upper. These are respectively: the required sample size, the lower bound of the range of values not detectable with the returned required sample size, calculated as baseline * (1 - mde), and the upper bound of the range of values not detectable with the returned required sample size, calculated as baseline * (1 + mde) (Float64). Tuple(Float64, Float64, Float64)

Examples

minSampleSizeContinuous

Query

SELECT minSampleSizeContinuous(112.25, 21.1, 0.03, 0.80, 0.05) AS sample_size

Response

(616.2931945826209,108.8825,115.6175)

minSampleSizeConversion

Introduced in: v22.6.0

Calculates minimum required sample size for an A/B test comparing conversions (proportions) in two samples.

Uses the formula described in this article. Assumes equal sizes of treatment and control groups. Returns the sample size required for one group (i.e. the sample size required for the whole experiment is twice the returned value).

Syntax

minSampleSizeConversion(baseline, mde, power, alpha)

Arguments

• baseline — Baseline conversion. Float*
• mde — Minimum detectable effect (MDE) as percentage points (e.g. for a baseline conversion 0.25 the MDE 0.03 means an expected change to 0.25 ± 0.03). Float*
• power — Required statistical power of a test (1 - probability of Type II error). Float*
• alpha — Required significance level of a test (probability of Type I error). Float*

Returned value

Returns a named Tuple with 3 elements: minimum_sample_size, detect_range_lower, detect_range_upper. These are, respectively: the required sample size, the lower bound of the range of values not detectable with the returned required sample size, calculated as baseline - mde, the upper bound of the range of values not detectable with the returned required sample size, calculated as baseline + mde. Tuple(Float64, Float64, Float64)

Examples

minSampleSizeConversion

Query

SELECT minSampleSizeConversion(0.25, 0.03, 0.80, 0.05) AS sample_size

Response

(3396.077603219163,0.22,0.28)

neighbor

Introduced in: v20.1.0

Returns a value from a column at a specified offset from the current row. This function is deprecated and error-prone because it operates on the physical order of data blocks which may not correspond to the logical order expected by users. Consider using proper window functions instead.

The function can be enabled by setting allow_deprecated_error_prone_window_functions = 1.

Syntax

neighbor(column, offset[, default_value])

Arguments

• column — The source column. Any
• offset — The offset from the current row. Positive values look forward, negative values look backward. Integer
• default_value — Optional. The value to return if the offset goes beyond the data bounds. If not specified, uses the default value for the column type. Any

Returned value

Returns a value from the specified offset, or default if out of bounds. Any

Examples

Usage example

Query

SELECT number, neighbor(number, 2) FROM system.numbers LIMIT 10;

Response

┌─number─┬─neighbor(number, 2)─┐
│      0 │                   2 │
│      1 │                   3 │
│      2 │                   4 │
│      3 │                   5 │
│      4 │                   6 │
│      5 │                   7 │
│      6 │                   8 │
│      7 │                   9 │
│      8 │                   0 │
│      9 │                   0 │
└────────┴─────────────────────┘

With default value

Query

SELECT number, neighbor(number, 2, 999) FROM system.numbers LIMIT 10;

Response

┌─number─┬─neighbor(number, 2, 999)─┐
│      0 │                        2 │
│      1 │                        3 │
│      2 │                        4 │
│      3 │                        5 │
│      4 │                        6 │
│      5 │                        7 │
│      6 │                        8 │
│      7 │                        9 │
│      8 │                      999 │
│      9 │                      999 │
└────────┴──────────────────────────┘

normalizeQuery

Introduced in: v20.8.0

Replaces literals, sequences of literals and complex aliases (containing whitespace, more than two digits or at least 36 bytes long such as UUIDs) with placeholder ?.

Syntax

normalizeQuery(x)

Arguments

• x — Sequence of characters. String

Returned value

Returns the given sequence of characters with placeholders. String

Examples

Usage example

Query

SELECT normalizeQuery('[1, 2, 3, x]') AS query

Response

┌─query────┐
│ [?.., x] │
└──────────┘

normalizeQueryKeepNames

Introduced in: v21.2.0

Replaces literals and sequences of literals with placeholder ? but does not replace complex aliases (containing whitespace, more than two digits or at least 36 bytes long such as UUIDs). This helps better analyze complex query logs.

Syntax

normalizeQueryKeepNames(x)

Arguments

• x — Sequence of characters. String

Returned value

Returns the given sequence of characters with placeholders. String

Examples

Usage example

Query

SELECT normalizeQuery('SELECT 1 AS aComplexName123'), normalizeQueryKeepNames('SELECT 1 AS aComplexName123')

Response

┌─normalizeQuery('SELECT 1 AS aComplexName123')─┬─normalizeQueryKeepNames('SELECT 1 AS aComplexName123')─┐
│ SELECT ? AS `?`                               │ SELECT ? AS aComplexName123                            │
└───────────────────────────────────────────────┴────────────────────────────────────────────────────────┘

normalizedQueryHash

Introduced in: v20.8.0

Returns identical 64 bit hash values without the values of literals for similar queries. Can be helpful in analyzing query logs.

Syntax

normalizedQueryHash(x)

Arguments

• x — Sequence of characters. String

Returned value

Returns a 64 bit hash value. UInt64

Examples

Usage example

Query

SELECT normalizedQueryHash('SELECT 1 AS `xyz`') != normalizedQueryHash('SELECT 1 AS `abc`') AS res

Response

┌─res─┐
│   1 │
└─────┘

normalizedQueryHashKeepNames

Introduced in: v21.2.0

Like normalizedQueryHash it returns identical 64 bit hash values without the values of literals for similar queries, but it does not replace complex aliases (containing whitespace, more than two digits or at least 36 bytes long such as UUIDs) with a placeholder before hashing. Can be helpful in analyzing query logs.

Syntax

normalizedQueryHashKeepNames(x)

Arguments

• x — Sequence of characters. String

Returned value

Returns a 64 bit hash value. UInt64

Examples

Usage example

Query

SELECT normalizedQueryHash('SELECT 1 AS `xyz123`') != normalizedQueryHash('SELECT 1 AS `abc123`') AS normalizedQueryHash;
SELECT normalizedQueryHashKeepNames('SELECT 1 AS `xyz123`') != normalizedQueryHashKeepNames('SELECT 1 AS `abc123`') AS normalizedQueryHashKeepNames;

Response

┌─normalizedQueryHash─┐
│                   0 │
└─────────────────────┘
┌─normalizedQueryHashKeepNames─┐
│                            1 │
└──────────────────────────────┘

obfuscateQuery

Introduced in: v26.4.0

Obfuscates a SQL query by replacing identifiers with random words and literals with random values while preserving query structure.

This function is useful for anonymizing queries before logging or sharing them for debugging purposes. Different rows will produce different obfuscated results even for the same input query, which helps maintain privacy when working with multiple queries.

The optional tag parameter prevents common subexpression elimination when the same function call is used multiple times in a query. This ensures that each invocation produces a different obfuscated result.

Features:

• Replaces table names, column names, and aliases with random words
• Replaces numeric and string literals with random values
• Preserves the overall query structure and SQL syntax
• Produces different results for different rows

Syntax

obfuscateQuery(query[, tag])

Arguments

• query — The SQL query to obfuscate. String
• tag — Optional. A value to prevent common subexpression elimination when the same function call is used multiple times.

Returned value

The obfuscated query with identifiers and literals replaced while preserving the original query structure. String

Examples

Basic usage

Query

SELECT obfuscateQuery('SELECT name, age FROM users WHERE age > 30')

Response

SELECT fruit, number FROM table WHERE number > 12

With tag to prevent common subexpression elimination

Query

SELECT obfuscateQuery('SELECT * FROM t', 1), obfuscateQuery('SELECT * FROM t', 2)

Response

SELECT a FROM b, SELECT c FROM d

Different rows produce different results

Query

SELECT obfuscateQuery('SELECT 1') AS a, obfuscateQuery('SELECT 1') AS b

Response

A B

obfuscateQueryWithSeed

Introduced in: v26.4.0

Obfuscates a SQL query using a specified seed for deterministic results.

Unlike obfuscateQuery(), this function produces deterministic results when given the same seed. This is useful when you need consistent obfuscation across multiple runs or when you want to reproduce the same obfuscated query for testing or debugging purposes.

Features:

• Deterministic obfuscation based on the provided seed
• Same seed always produces the same obfuscated result
• Different seeds produce different results
• Preserves query structure like obfuscateQuery()

Use cases:

• Reproducible test cases
• Consistent anonymization across multiple runs
• Debugging with consistent obfuscated queries

Syntax

obfuscateQueryWithSeed(query, seed)

Arguments

• query — The SQL query to obfuscate. String
• seed — The seed for obfuscation. The same seed produces deterministic results. Integer or String

Returned value

The obfuscated query, deterministically generated based on the provided seed. String

Examples

Deterministic obfuscation with integer seed

Query

SELECT obfuscateQueryWithSeed('SELECT name FROM users', 42)

Response

SELECT fruit FROM table

Deterministic obfuscation with string seed

Query

SELECT obfuscateQueryWithSeed('SELECT id, value FROM data', 'myseed')

Response

SELECT a, b FROM c

Same seed produces same result

Query

SELECT obfuscateQueryWithSeed('SELECT 1', 100) = obfuscateQueryWithSeed('SELECT 1', 100)

Response

true

parseReadableSize

Introduced in: v24.6.0

Given a string containing a byte size and B, KiB, KB, MiB, MB, etc. as a unit (i.e. ISO/IEC 80000-13 or decimal byte unit), this function returns the corresponding number of bytes. If the function is unable to parse the input value, it throws an exception.

The inverse operations of this function are formatReadableSize and formatReadableDecimalSize.

Syntax

parseReadableSize(x)

Arguments

• x — Readable size with ISO/IEC 80000-13 or decimal byte unit. String

Returned value

Returns the number of bytes, rounded up to the nearest integer. UInt64

Examples

Usage example

Query

SELECT arrayJoin(['1 B', '1 KiB', '3 MB', '5.314 KiB']) AS readable_sizes, parseReadableSize(readable_sizes) AS sizes;

Response

┌─readable_sizes─┬───sizes─┐
│ 1 B            │       1 │
│ 1 KiB          │    1024 │
│ 3 MB           │ 3000000 │
│ 5.314 KiB      │    5442 │
└────────────────┴─────────┘

parseReadableSizeOrNull

Introduced in: v24.6.0

Given a string containing a byte size and B, KiB, KB, MiB, MB, etc. as a unit (i.e. ISO/IEC 80000-13 or decimal byte unit), this function returns the corresponding number of bytes. If the function is unable to parse the input value, it returns NULL.

The inverse operations of this function are formatReadableSize and formatReadableDecimalSize.

Syntax

parseReadableSizeOrNull(x)

Arguments

• x — Readable size with ISO/IEC 80000-13 or decimal byte unit. String

Returned value

Returns the number of bytes, rounded up to the nearest integer, or NULL if unable to parse the input Nullable(UInt64)

Examples

Usage example

Query

SELECT arrayJoin(['1 B', '1 KiB', '3 MB', '5.314 KiB', 'invalid']) AS readable_sizes, parseReadableSizeOrNull(readable_sizes) AS sizes;

Response

┌─readable_sizes─┬───sizes─┐
│ 1 B            │       1 │
│ 1 KiB          │    1024 │
│ 3 MB           │ 3000000 │
│ 5.314 KiB      │    5442 │
│ invalid        │    ᴺᵁᴸᴸ │
└────────────────┴─────────┘

parseReadableSizeOrZero

Introduced in: v24.6.0

Given a string containing a byte size and B, KiB, KB, MiB, MB, etc. as a unit (i.e. ISO/IEC 80000-13 or decimal byte unit), this function returns the corresponding number of bytes. If the function is unable to parse the input value, it returns 0.

The inverse operations of this function are formatReadableSize and formatReadableDecimalSize.

Syntax

parseReadableSizeOrZero(x)

Arguments

• x — Readable size with ISO/IEC 80000-13 or decimal byte unit. String

Returned value

Returns the number of bytes, rounded up to the nearest integer, or 0 if unable to parse the input. UInt64

Examples

Usage example

Query

SELECT arrayJoin(['1 B', '1 KiB', '3 MB', '5.314 KiB', 'invalid']) AS readable_sizes, parseReadableSizeOrZero(readable_sizes) AS sizes;

Response

┌─readable_sizes─┬───sizes─┐
│ 1 B            │       1 │
│ 1 KiB          │    1024 │
│ 3 MB           │ 3000000 │
│ 5.314 KiB      │    5442 │
│ invalid        │       0 │
└────────────────┴─────────┘

parseTimeDelta

Introduced in: v22.7.0

Parse a sequence of numbers followed by something resembling a time unit.

The time delta string uses these time unit specifications:

• years, year, yr, y
• months, month, mo
• weeks, week, w
• days, day, d
• hours, hour, hr, h
• minutes, minute, min, m
• seconds, second, sec, s
• milliseconds, millisecond, millisec, ms
• microseconds, microsecond, microsec, μs, µs, us
• nanoseconds, nanosecond, nanosec, ns

Multiple time units can be combined with separators (space, ;, -, +, ,, :).

The length of years and months are approximations: year is 365 days, month is 30.5 days.

Syntax

parseTimeDelta(timestr)

Arguments

• timestr — A sequence of numbers followed by something resembling a time unit. String

Returned value

The number of seconds. Float64

Examples

Usage example

Query

SELECT parseTimeDelta('11s+22min')

Response

┌─parseTimeDelta('11s+22min')─┐
│                        1331 │
└─────────────────────────────┘

Complex time units

Query

SELECT parseTimeDelta('1yr2mo')

Response

┌─parseTimeDelta('1yr2mo')─┐
│                 36806400 │
└──────────────────────────┘

partitionId

Introduced in: v21.4.0

Computes the partition ID.

This function is slow and should not be called for large numbers of rows.

Syntax

partitionId(column1[, column2, ...])

Aliases: partitionID

Arguments

• column1, column2, ... — Column for which to return the partition ID.

Returned value

Returns the partition ID that the row belongs to. String

Examples

Usage example

Query

DROP TABLE IF EXISTS tab;

CREATE TABLE tab
(
  i int,
  j int
)
ENGINE = MergeTree
PARTITION BY i
ORDER BY tuple();

INSERT INTO tab VALUES (1, 1), (1, 2), (1, 3), (2, 4), (2, 5), (2, 6);

SELECT i, j, partitionId(i), _partition_id FROM tab ORDER BY i, j;

Response

┌─i─┬─j─┬─partitionId(i)─┬─_partition_id─┐
│ 1 │ 1 │ 1              │ 1             │
│ 1 │ 2 │ 1              │ 1             │
│ 1 │ 3 │ 1              │ 1             │
│ 2 │ 4 │ 2              │ 2             │
│ 2 │ 5 │ 2              │ 2             │
│ 2 │ 6 │ 2              │ 2             │
└───┴───┴────────────────┴───────────────┘

queryID

Introduced in: v21.9.0

Returns the ID of the current query. Other parameters of a query can be extracted from field query_id in the system.query_log table.

In contrast to initialQueryID function, queryID can return different results on different shards.

Syntax

queryID()

Aliases: query_id

Arguments

• None.

Returned value

Returns the ID of the current query. String

Examples

Usage example

Query

CREATE TABLE tmp (str String) ENGINE = Log;
INSERT INTO tmp (*) VALUES ('a');
SELECT count(DISTINCT t) FROM (SELECT queryID() AS t FROM remote('127.0.0.{1..3}', currentDatabase(), 'tmp') GROUP BY queryID());

Response

┌─count(DISTINCT t)─┐
│                 3 │
└───────────────────┘

revision

Introduced in: v22.7.0

Returns the current ClickHouse server revision.

Syntax

revision()

Arguments

• None.

Returned value

Returns the current ClickHouse server revision. UInt32

Examples

Usage example

Query

SELECT revision()

Response

┌─revision()─┐
│      54485 │
└────────────┘

rowNumberInAllBlocks

Introduced in: v1.1.0

Returns a unique row number for each row processed.

Syntax

rowNumberInAllBlocks()

Arguments

• None.

Returned value

Returns the ordinal number of the row in the data block starting from 0. UInt64

Examples

Usage example

Query

SELECT rowNumberInAllBlocks()
FROM
(
    SELECT *
    FROM system.numbers_mt
    LIMIT 10
)
SETTINGS max_block_size = 2

Response

┌─rowNumberInAllBlocks()─┐
│                      0 │
│                      1 │
└────────────────────────┘
┌─rowNumberInAllBlocks()─┐
│                      4 │
│                      5 │
└────────────────────────┘
┌─rowNumberInAllBlocks()─┐
│                      2 │
│                      3 │
└────────────────────────┘
┌─rowNumberInAllBlocks()─┐
│                      6 │
│                      7 │
└────────────────────────┘
┌─rowNumberInAllBlocks()─┐
│                      8 │
│                      9 │
└────────────────────────┘

rowNumberInBlock

Introduced in: v1.1.0

For each block processed by rowNumberInBlock, returns the number of the current row.

The returned number starts from 0 for each block.

Syntax

rowNumberInBlock()

Arguments

• None.

Returned value

Returns the ordinal number of the row in the data block starting from 0. UInt64

Examples

Usage example

Query

SELECT rowNumberInBlock()
FROM
(
    SELECT *
    FROM system.numbers_mt
    LIMIT 10
) SETTINGS max_block_size = 2

Response

┌─rowNumberInBlock()─┐
│                  0 │
│                  1 │
└────────────────────┘
┌─rowNumberInBlock()─┐
│                  0 │
│                  1 │
└────────────────────┘
┌─rowNumberInBlock()─┐
│                  0 │
│                  1 │
└────────────────────┘
┌─rowNumberInBlock()─┐
│                  0 │
│                  1 │
└────────────────────┘
┌─rowNumberInBlock()─┐
│                  0 │
│                  1 │
└────────────────────┘

runningAccumulate

Introduced in: v1.1.0

Accumulates the states of an aggregate function for each row of a data block.

Deprecated

The state is reset for each new block of data. Due to this error-prone behavior the function has been deprecated, and you are advised to use window functions instead. You can use setting allow_deprecated_error_prone_window_functions to allow usage of this function.

Syntax

runningAccumulate(agg_state[, grouping])

Arguments

• agg_state — State of the aggregate function. AggregateFunction
• grouping — Optional. Grouping key. The state of the function is reset if the grouping value is changed. It can be any of the supported data types for which the equality operator is defined. Any

Returned value

Returns the accumulated result for each row. Any

Examples

Usage example with initializeAggregation

Query

WITH initializeAggregation('sumState', number) AS one_row_sum_state
SELECT
    number,
    finalizeAggregation(one_row_sum_state) AS one_row_sum,
    runningAccumulate(one_row_sum_state) AS cumulative_sum
FROM numbers(5);

Response

┌─number─┬─one_row_sum─┬─cumulative_sum─┐
│      0 │           0 │              0 │
│      1 │           1 │              1 │
│      2 │           2 │              3 │
│      3 │           3 │              6 │
│      4 │           4 │             10 │
└────────┴─────────────┴────────────────┘

runningConcurrency

Introduced in: v21.3.0

Calculates the number of concurrent events. Each event has a start time and an end time. The start time is included in the event, while the end time is excluded. Columns with a start time and an end time must be of the same data type. The function calculates the total number of active (concurrent) events for each event start time.

Requirements

Events must be ordered by the start time in ascending order. If this requirement is violated the function raises an exception. Every data block is processed separately. If events from different data blocks overlap then they can not be processed correctly.

Deprecated

It is advised to use window functions instead.

Syntax

runningConcurrency(start, end)

Arguments

• start — A column with the start time of events. Date or DateTime or DateTime64
• end — A column with the end time of events. Date or DateTime or DateTime64

Returned value

Returns the number of concurrent events at each event start time. UInt32

Examples

Usage example

Query

SELECT start, runningConcurrency(start, end) FROM example_table;

Response

┌──────start─┬─runningConcurrency(start, end)─┐
│ 2025-03-03 │                              1 │
│ 2025-03-06 │                              2 │
│ 2025-03-07 │                              3 │
│ 2025-03-11 │                              2 │
└────────────┴────────────────────────────────┘

runningDifference

Introduced in: v1.1.0

Calculates the difference between two consecutive row values in the data block. Returns 0 for the first row, and for subsequent rows the difference to the previous row.

Deprecated

Only returns differences inside the currently processed data block. Because of this error-prone behavior, the function is deprecated. It is advised to use window functions instead.

You can use setting allow_deprecated_error_prone_window_functions to allow usage of this function.

The result of the function depends on the affected data blocks and the order of data in the block. The order of rows during calculation of runningDifference() can differ from the order of rows returned to the user. To prevent that you can create a subquery with ORDER BY and call the function from outside the subquery. Please note that the block size affects the result. The internal state of runningDifference state is reset for each new block.

Syntax

runningDifference(x)

Arguments

• x — Column for which to calculate the running difference. Any

Returned value

Returns the difference between consecutive values, with 0 for the first row.

Examples

Usage example

Query

SELECT
    EventID,
    EventTime,
    runningDifference(EventTime) AS delta
FROM
(
    SELECT
        EventID,
        EventTime
    FROM events
    WHERE EventDate = '2025-11-24'
    ORDER BY EventTime ASC
    LIMIT 5
);

Response

┌─EventID─┬───────────EventTime─┬─delta─┐
│    1106 │ 2025-11-24 00:00:04 │     0 │
│    1107 │ 2025-11-24 00:00:05 │     1 │
│    1108 │ 2025-11-24 00:00:05 │     0 │
│    1109 │ 2025-11-24 00:00:09 │     4 │
│    1110 │ 2025-11-24 00:00:10 │     1 │
└─────────┴─────────────────────┴───────┘

Block size impact example

Query

SELECT
    number,
    runningDifference(number + 1) AS diff
FROM numbers(100000)
WHERE diff != 1;

Response

┌─number─┬─diff─┐
│      0 │    0 │
└────────┴──────┘
┌─number─┬─diff─┐
│  65536 │    0 │
└────────┴──────┘

runningDifferenceStartingWithFirstValue

Introduced in: v1.1.0

Calculates the difference between consecutive row values in a data block, but unlike runningDifference, it returns the actual value of the first row instead of 0.

Deprecated

Only returns differences inside the currently processed data block. Because of this error-prone behavior, the function is deprecated. It is advised to use window functions instead.

You can use setting allow_deprecated_error_prone_window_functions to allow usage of this function.

Syntax

runningDifferenceStartingWithFirstValue(x)

Arguments

• x — Column for which to calculate the running difference. Any

Returned value

Returns the difference between consecutive values, with the first row’s value for the first row. Any

Examples

Usage example

Query

SELECT
    number,
    runningDifferenceStartingWithFirstValue(number) AS diff
FROM numbers(5);

Response

┌─number─┬─diff─┐
│      0 │    0 │
│      1 │    1 │
│      2 │    1 │
│      3 │    1 │
│      4 │    1 │
└────────┴──────┘

serverUUID

Introduced in: v20.1.0

Returns the random and unique UUID (v4) generated when the server is first started. The UUID is persisted, i.e. the second, third, etc. server start return the same UUID.

Syntax

serverUUID()

Arguments

• None.

Returned value

Returns the random UUID of the server. UUID

Examples

Usage example

Query

SELECT serverUUID();

Response

┌─serverUUID()─────────────────────────────┐
│ 7ccc9260-000d-4d5c-a843-5459abaabb5f     │
└──────────────────────────────────────────┘

shardCount

Introduced in: v21.9.0

Returns the total number of shards for a distributed query. If a query is not distributed then constant value 0 is returned.

Syntax

shardCount()

Arguments

• None.

Returned value

Returns the total number of shards or 0. UInt32

Examples

Usage example

Query

-- See shardNum() example above which also demonstrates shardCount()
CREATE TABLE shard_count_example (dummy UInt8)
ENGINE=Distributed(test_cluster_two_shards_localhost, system, one, dummy);
SELECT shardCount() FROM shard_count_example;

Response

┌─shardCount()─┐
│            2 │
│            2 │
└──────────────┘

shardNum

Introduced in: v21.9.0

Returns the index of a shard which processes a part of data in a distributed query. Indices begin from 1. If a query is not distributed then a constant value 0 is returned.

Syntax

shardNum()

Arguments

• None.

Returned value

Returns the shard index or a constant 0. UInt32

Examples

Usage example

Query

CREATE TABLE shard_num_example (dummy UInt8)
ENGINE=Distributed(test_cluster_two_shards_localhost, system, one, dummy);
SELECT dummy, shardNum(), shardCount() FROM shard_num_example;

Response

┌─dummy─┬─shardNum()─┬─shardCount()─┐
│     0 │          1 │            2 │
│     0 │          2 │            2 │
└───────┴────────────┴──────────────┘

showCertificate

Introduced in: v22.6.0

Shows information about the current server’s Secure Sockets Layer (SSL) certificate if it has been configured. See Configuring TLS for more information on how to configure ClickHouse to use OpenSSL certificates to validate connections.

Syntax

showCertificate()

Arguments

• None.

Returned value

Returns map of key-value pairs relating to the configured SSL certificate. Map(String, String)

Examples

Usage example

Query

SELECT showCertificate() FORMAT LineAsString;

Response

{'version':'1','serial_number':'2D9071D64530052D48308473922C7ADAFA85D6C5','signature_algo':'sha256WithRSAEncryption','issuer':'/CN=marsnet.local CA','not_before':'May  7 17:01:21 2024 GMT','not_after':'May  7 17:01:21 2025 GMT','subject':'/CN=chnode1','pkey_algo':'rsaEncryption'}

sleep

Introduced in: v1.1.0

Pauses the execution of a query by the specified number of seconds. The function is primarily used for testing and debugging purposes.

The sleep() function should generally not be used in production environments, as it can negatively impact query performance and system responsiveness. However, it can be useful in the following scenarios:

1. Testing: When testing or benchmarking ClickHouse, you may want to simulate delays or introduce pauses to observe how the system behaves under certain conditions.
2. Debugging: If you need to examine the state of the system or the execution of a query at a specific point in time, you can use sleep() to introduce a pause, allowing you to inspect or collect relevant information.
3. Simulation: In some cases, you may want to simulate real-world scenarios where delays or pauses occur, such as network latency or external system dependencies.

It’s important to use the sleep() function judiciously and only when necessary, as it can potentially impact the overall performance and responsiveness of your ClickHouse system.

For security reasons, the function can only be executed in the default user profile (with allow_sleep enabled).

Syntax

sleep(seconds)

Arguments

• seconds — The number of seconds to pause the query execution to a maximum of 3 seconds. It can be a floating-point value to specify fractional seconds. const UInt* or const Float*

Returned value

Returns 0. UInt8

Examples

Usage example

Query

-- This query will pause for 2 seconds before completing.
-- During this time, no results will be returned, and the query will appear to be hanging or unresponsive.
SELECT sleep(2);

Response

┌─sleep(2)─┐
│        0 │
└──────────┘
1 row in set. Elapsed: 2.012 sec.

sleepEachRow

Introduced in: v1.1.0

Pauses the execution of a query for a specified number of seconds for each row in the result set.

The sleepEachRow() function is primarily used for testing and debugging purposes, similar to the sleep() function. It allows you to simulate delays or introduce pauses in the processing of each row, which can be useful in scenarios such as:

1. Testing: When testing or benchmarking ClickHouse’s performance under specific conditions, you can use sleepEachRow() to simulate delays or introduce pauses for each row processed.
2. Debugging: If you need to examine the state of the system or the execution of a query for each row processed, you can use sleepEachRow() to introduce pauses, allowing you to inspect or collect relevant information.
3. Simulation: In some cases, you may want to simulate real-world scenarios where delays or pauses occur for each row processed, such as when dealing with external systems or network latencies.

Like the sleep() function, it’s important to use sleepEachRow() judiciously and only when necessary, as it can significantly impact the overall performance and responsiveness of your ClickHouse system, especially when dealing with large result sets.

Syntax

sleepEachRow(seconds)

Arguments

• seconds — The number of seconds to pause the query execution for each row in the result set to a maximum of 3 seconds. It can be a floating-point value to specify fractional seconds. const UInt* or const Float*

Returned value

Returns 0 for each row. UInt8

Examples

Usage example

Query

-- The output will be delayed, with a 0.5-second pause between each row.
SELECT number, sleepEachRow(0.5) FROM system.numbers LIMIT 5;

Response

┌─number─┬─sleepEachRow(0.5)─┐
│      0 │                 0 │
│      1 │                 0 │
│      2 │                 0 │
│      3 │                 0 │
│      4 │                 0 │
└────────┴───────────────────┘

structureToCapnProtoSchema

Introduced in: v23.8.0

Function that converts ClickHouse table structure to CapnProto format schema

Syntax

structureToCapnProtoSchema(table_structure, message)

Arguments

• None.

Returned value

Examples

random

Query

SELECT structureToCapnProtoSchema('s String, x UInt32', 'MessageName') format TSVRaw

Response

struct MessageName
{
    s @0 : Data;
    x @1 : UInt32;
}

structureToProtobufSchema

Introduced in: v23.8.0

Converts a ClickHouse table structure to Protobuf format schema.

This function takes a ClickHouse table structure definition and converts it into a Protocol Buffers (Protobuf) schema definition in proto3 syntax. This is useful for generating Protobuf schemas that match your ClickHouse table structures for data interchange.

Syntax

structureToProtobufSchema(structure, message_name)

Arguments

• structure — ClickHouse table structure definition as a string (e.g., ‘column1 Type1, column2 Type2’). String
• message_name — Name for the Protobuf message type in the generated schema. String

Returned value

Returns a Protobuf schema definition in proto3 syntax that corresponds to the input ClickHouse structure. String

Examples

Converting ClickHouse structure to Protobuf schema

Query

SELECT structureToProtobufSchema('s String, x UInt32', 'MessageName') FORMAT TSVRaw;

Response

syntax = "proto3";

message MessageName
{
    bytes s = 1;
    uint32 x = 2;
}

tcpPort

Introduced in: v20.12.0

Returns the native interface TCP port number listened to by the server. If executed in the context of a distributed table, this function generates a normal column with values relevant to each shard. Otherwise it produces a constant value.

Syntax

tcpPort()

Arguments

• None.

Returned value

Returns the TCP port number. UInt16

Examples

Usage example

Query

SELECT tcpPort()

Response

┌─tcpPort()─┐
│      9000 │
└───────────┘

throwIf

Introduced in: v1.1.0

Throw an exception if argument x is true. To use the error_code argument, configuration parameter allow_custom_error_code_in_throw must be enabled.

Syntax

throwIf(x[, message[, error_code]])

Arguments

• x — The condition to check. Any
• message — Optional. Custom error message. const String
• error_code — Optional. Custom error code. const Int8/16/32

Returned value

Returns 0 if the condition is false, throws an exception if the condition is true. UInt8

Examples

Usage example

Query

SELECT throwIf(number = 3, 'Too many') FROM numbers(10);

Response

↙ Progress: 0.00 rows, 0.00 B (0.00 rows/s., 0.00 B/s.) Received exception from server (version 19.14.1):
Code: 395. DB::Exception: Received from localhost:9000. DB::Exception: Too many.

toColumnTypeName

Introduced in: v1.1.0

Returns the internal name of the data type of the given value. Unlike function toTypeName, the returned data type potentially includes internal wrapper columns like Const and LowCardinality.

Syntax

toColumnTypeName(value)

Arguments

• value — Value for which to return the internal data type. Any

Returned value

Returns the internal data type used to represent the value. String

Examples

Usage example

Query

SELECT toColumnTypeName(CAST('2025-01-01 01:02:03' AS DateTime));

Response

┌─toColumnTypeName(CAST('2025-01-01 01:02:03', 'DateTime'))─┐
│ Const(UInt32)                                             │
└───────────────────────────────────────────────────────────┘

toTypeName

Introduced in: v1.1.0

Returns the type name of the passed argument. If NULL is passed, the function returns type Nullable(Nothing), which corresponds to ClickHouse’s internal NULL representation.

Syntax

toTypeName(x)

Arguments

• x — A value of arbitrary type. Any

Returned value

Returns the data type name of the input value. String

Examples

Usage example

Query

SELECT toTypeName(123)

Response

┌─toTypeName(123)─┐
│ UInt8           │
└─────────────────┘

tokenizeQuery

Introduced in: v26.5.0

Tokenizes a ClickHouse SQL query string and returns an array of tokens. Each token is a named tuple with the beginning position (in bytes), the end position, and the token type.

Syntax

tokenizeQuery(query)

Arguments

• query — A ClickHouse SQL query string. String.

Returned value

An array of named tuples (begin UInt64, end UInt64, type Enum8(...)) representing the tokens of the query. Array(Tuple(begin UInt64, end UInt64, type Enum8(...)))

Examples

simple

Query

SELECT tokenizeQuery('SELECT 1')

Response

[(0,6,'BareWord'),(6,7,'Whitespace'),(7,8,'Number')]

transactionID

Introduced in: v22.6.0

Returns the ID of a transaction.

This function is part of an experimental feature set. Enable experimental transaction support by adding this setting to your configuration:

<clickhouse>
    <allow_experimental_transactions>1</allow_experimental_transactions>
</clickhouse>

For more information see the page Transactional (ACID) support.

Syntax

transactionID()

Arguments

• None.

Returned value

Returns a tuple consisting of start_csn, local_tid and host_id.

• start_csn: Global sequential number, the newest commit timestamp that was seen when this transaction began.
• local_tid: Local sequential number that is unique for each transaction started by this host within a specific start_csn.
• host_id: UUID of the host that has started this transaction. Tuple(UInt64, UInt64, UUID)

Examples

Usage example

Query

BEGIN TRANSACTION;
SELECT transactionID();
ROLLBACK;

Response

┌─transactionID()────────────────────────────────┐
│ (32,34,'0ee8b069-f2bb-4748-9eae-069c85b5252b') │
└────────────────────────────────────────────────┘

transactionLatestSnapshot

Introduced in: v22.6.0

Returns the newest snapshot (Commit Sequence Number) of a transaction that is available for reading.

This function is part of an experimental feature set. Enable experimental transaction support by adding this setting to your configuration:

<clickhouse>
    <allow_experimental_transactions>1</allow_experimental_transactions>
</clickhouse>

For more information see the page Transactional (ACID) support.

Syntax

transactionLatestSnapshot()

Arguments

• None.

Returned value

Returns the latest snapshot (CSN) of a transaction. UInt64

Examples

Usage example

Query

BEGIN TRANSACTION;
SELECT transactionLatestSnapshot();
ROLLBACK;

Response

┌─transactionLatestSnapshot()─┐
│                          32 │
└─────────────────────────────┘

transactionOldestSnapshot

Introduced in: v22.6.0

Returns the oldest snapshot (Commit Sequence Number) that is visible for some running transaction.

This function is part of an experimental feature set. Enable experimental transaction support by adding this setting to your configuration:

<clickhouse>
    <allow_experimental_transactions>1</allow_experimental_transactions>
</clickhouse>

For more information see the page Transactional (ACID) support.

Syntax

transactionOldestSnapshot()

Arguments

• None.

Returned value

Returns the oldest snapshot (CSN) of a transaction. UInt64

Examples

Usage example

Query

BEGIN TRANSACTION;
SELECT transactionOldestSnapshot();
ROLLBACK;

Response

┌─transactionOldestSnapshot()─┐
│                          32 │
└─────────────────────────────┘

transform

Introduced in: v1.1.0

Transforms a value according to the explicitly defined mapping of some elements to other elements.

There are two variations of this function:

• transform(x, array_from, array_to, default) - transforms x using mapping arrays with a default value for unmatched elements
• transform(x, array_from, array_to) - same transformation but returns the original x if no match is found

The function searches for x in array_from and returns the corresponding element from array_to at the same index. If x is not found in array_from, it returns either the default value (4-parameter version) or the original x (3-parameter version). If multiple matching elements exist in array_from, it returns the element corresponding to the first match.

Requirements:

• array_from and array_to must have the same number of elements
• For 4-parameter version: transform(T, Array(T), Array(U), U) -> U where T and U can be different compatible types
• For 3-parameter version: transform(T, Array(T), Array(T)) -> T where all types must be the same

Syntax

transform(x, array_from, array_to[, default])

Arguments

• x — Value to transform. (U)Int* or Decimal or Float* or String or Date or DateTime
• array_from — Constant array of values to search for matches. Array((U)Int*) or Array(Decimal) or Array(Float*) or Array(String) or Array(Date) or Array(DateTime)
• array_to — Constant array of values to return for corresponding matches in array_from. Array((U)Int*) or Array(Decimal) or Array(Float*) or Array(String) or Array(Date) or Array(DateTime)
• default — Optional. Value to return if x is not found in array_from. If omitted, returns x unchanged. (U)Int* or Decimal or Float* or String or Date or DateTime

Returned value

Returns the corresponding value from array_to if x matches an element in array_from, otherwise returns default (if provided) or x (if default not provided). Any

Examples

transform(T, Array(T), Array(U), U) -> U

Query

SELECT
transform(SearchEngineID, [2, 3], ['Yandex', 'Google'], 'Other') AS title,
count() AS c
FROM test.hits
WHERE SearchEngineID != 0
GROUP BY title
ORDER BY c DESC

Response

┌─title─────┬──────c─┐
│ Yandex    │ 498635 │
│ Google    │ 229872 │
│ Other     │ 104472 │
└───────────┴────────┘

transform(T, Array(T), Array(T)) -> T

Query

SELECT
transform(domain(Referer), ['yandex.ru', 'google.ru', 'vkontakte.ru'], ['www.yandex', 'example.com', 'vk.com']) AS s, count() AS c
FROM test.hits
GROUP BY domain(Referer)
ORDER BY count() DESC
LIMIT 10

Response

┌─s──────────────┬───────c─┐
│                │ 2906259 │
│ www.yandex     │  867767 │
│ ███████.ru     │  313599 │
│ mail.yandex.ru │  107147 │
│ ██████.ru      │  100355 │
│ █████████.ru   │   65040 │
│ news.yandex.ru │   64515 │
│ ██████.net     │   59141 │
│ example.com    │   57316 │
└────────────────┴─────────┘

uniqThetaIntersect

Introduced in: v22.9.0

Two uniqThetaSketch objects to do intersect calculation(set operation ∩), the result is a new uniqThetaSketch.

Syntax

uniqThetaIntersect(uniqThetaSketch,uniqThetaSketch)

Arguments

• uniqThetaSketch — uniqThetaSketch object. Tuple or Array or Date or DateTime or String or (U)Int* or Float* or Decimal

Returned value

A new uniqThetaSketch containing the intersect result. UInt64

Examples

Usage example

Query

SELECT finalizeAggregation(uniqThetaIntersect(a, b)) AS a_intersect_b, finalizeAggregation(a) AS a_cardinality, finalizeAggregation(b) AS b_cardinality
FROM
(SELECT arrayReduce('uniqThetaState', [1, 2]) AS a, arrayReduce('uniqThetaState', [2, 3, 4]) AS b);

Response

┌─a_intersect_b─┬─a_cardinality─┬─b_cardinality─┐
│             1 │             2 │             3 │
└───────────────┴───────────────┴───────────────┘

uniqThetaNot

Introduced in: v22.9.0

Two uniqThetaSketch objects to do a_not_b calculation(set operation ×), the result is a new uniqThetaSketch.

Syntax

uniqThetaNot(uniqThetaSketch,uniqThetaSketch)

Arguments

• uniqThetaSketch — uniqThetaSketch object. Tuple or Array or Date or DateTime or String or (U)Int* or Float* or Decimal

Returned value

Returns a new uniqThetaSketch containing the a_not_b result. UInt64

Examples

Usage example

Query

SELECT finalizeAggregation(uniqThetaNot(a, b)) AS a_not_b, finalizeAggregation(a) AS a_cardinality, finalizeAggregation(b) AS b_cardinality
FROM
(SELECT arrayReduce('uniqThetaState', [2, 3, 4]) AS a, arrayReduce('uniqThetaState', [1, 2]) AS b);

Response

┌─a_not_b─┬─a_cardinality─┬─b_cardinality─┐
│       2 │             3 │             2 │
└─────────┴───────────────┴───────────────┘

uniqThetaUnion

Introduced in: v22.9.0

Two uniqThetaSketch objects to do union calculation(set operation ∪), the result is a new uniqThetaSketch.

Syntax

uniqThetaUnion(uniqThetaSketch,uniqThetaSketch)

Arguments

• uniqThetaSketch — uniqThetaSketch object. Tuple or Array or Date or DateTime or String or (U)Int* or Float* or Decimal

Returned value

Returns a new uniqThetaSketch containing the union result. UInt64

Examples

Usage example

Query

SELECT finalizeAggregation(uniqThetaUnion(a, b)) AS a_union_b, finalizeAggregation(a) AS a_cardinality, finalizeAggregation(b) AS b_cardinality
FROM
(SELECT arrayReduce('uniqThetaState', [1, 2]) AS a, arrayReduce('uniqThetaState', [2, 3, 4]) AS b);

Response

┌─a_union_b─┬─a_cardinality─┬─b_cardinality─┐
│         4 │             2 │             3 │
└───────────┴───────────────┴───────────────┘

uptime

Introduced in: v1.1.0

Returns the server’s uptime in seconds. If executed in the context of a distributed table, this function generates a normal column with values relevant to each shard. Otherwise it produces a constant value.

Syntax

uptime()

Arguments

• None.

Returned value

Returns the server uptime in seconds. UInt32

Examples

Usage example

Query

SELECT uptime() AS Uptime

Response

┌─Uptime─┐
│  55867 │
└────────┘

variantElement

Introduced in: v25.2.0

Extracts a column with specified type from a Variant column.

Syntax

variantElement(variant, type_name[, default_value])

Arguments

• variant — Variant column. Variant
• type_name — The name of the variant type to extract. String
• default_value — The default value that will be used if variant doesn’t have variant with specified type. Can be any type. Optional. Any

Returned value

Returns a column with the specified variant type extracted from the Variant column. Any

Examples

Usage example

Query

CREATE TABLE test (v Variant(UInt64, String, Array(UInt64))) ENGINE = Memory;
INSERT INTO test VALUES (NULL), (42), ('Hello, World!'), ([1, 2, 3]);
SELECT v, variantElement(v, 'String'), variantElement(v, 'UInt64'), variantElement(v, 'Array(UInt64)') FROM test;

Response

┌─v─────────────┬─variantElement(v, 'String')─┬─variantElement(v, 'UInt64')─┬─variantElement(v, 'Array(UInt64)')─┐
│ ᴺᵁᴸᴸ          │ ᴺᵁᴸᴸ                        │                        ᴺᵁᴸᴸ │ []                                 │
│ 42            │ ᴺᵁᴸᴸ                        │                          42 │ []                                 │
│ Hello, World! │ Hello, World!               │                        ᴺᵁᴸᴸ │ []                                 │
│ [1,2,3]       │ ᴺᵁᴸᴸ                        │                        ᴺᵁᴸᴸ │ [1,2,3]                            │
└───────────────┴─────────────────────────────┴─────────────────────────────┴────────────────────────────────────┘

variantType

Introduced in: v24.2.0

Returns the variant type name for each row of Variant column. If row contains NULL, it returns ‘None’ for it.

Syntax

variantType(variant)

Arguments

• variant — Variant column. Variant

Returned value

Returns an Enum column with variant type name for each row. Enum

Examples

Usage example

Query

CREATE TABLE test (v Variant(UInt64, String, Array(UInt64))) ENGINE = Memory;
INSERT INTO test VALUES (NULL), (42), ('Hello, World!'), ([1, 2, 3]);
SELECT variantType(v) FROM test;

Response

┌─variantType(v)─┐
│ None           │
│ UInt64         │
│ String         │
│ Array(UInt64)  │
└────────────────┘

version

Introduced in: v1.1.0

Returns the current version of ClickHouse as a string in the form: major_version.minor_version.patch_version.number_of_commits_since_the_previous_stable_release. If executed in the context of a distributed table, this function generates a normal column with values relevant to each shard. Otherwise, it produces a constant value.

Syntax

version()

Arguments

• None.

Returned value

Returns the current version of ClickHouse. String

Examples

Usage example

Query

SELECT version()

Response

┌─version()─┐
│ 24.2.1.1  │
└───────────┘

visibleWidth

Introduced in: v1.1.0

Calculates the approximate width when outputting values to the console in text format (tab-separated). This function is used by the system to implement Pretty formats. NULL is represented as a string corresponding to NULL in Pretty formats.

Syntax

visibleWidth(x)

Arguments

• x — A value of any data type. Any

Returned value

Returns the approximate width of the value when displayed in text format. UInt64

Examples

Calculate visible width of NULL

Query

SELECT visibleWidth(NULL)

Response

┌─visibleWidth(NULL)─┐
│                  4 │
└────────────────────┘

zookeeperSessionUptime

Introduced in: v21.11.0

Returns the uptime of the current ZooKeeper session in seconds.

Syntax

zookeeperSessionUptime()

Arguments

• None.

Returned value

Returns the uptime of the current ZooKeeper session in seconds. UInt32

Examples

Usage example

Query

SELECT zookeeperSessionUptime();

Response

┌─zookeeperSessionUptime()─┐
│                      286 │
└──────────────────────────┘