Hierarchical dictionaries

Configure hierarchical dictionaries with parent-child key relationships.

Hierarchical dictionaries

ClickHouse supports hierarchical dictionaries with a numeric key.

Look at the following hierarchical structure:

0 (Common parent)
│
├── 1 (Russia)
│   │
│   └── 2 (Moscow)
│       │
│       └── 3 (Center)
│
└── 4 (Great Britain)
    │
    └── 5 (London)

This hierarchy can be expressed as the following dictionary table.

region_id	parent_region	region_name
1	0	Russia
2	1	Moscow
3	2	Center
4	0	Great Britain
5	4	London

This table contains a column parent_region that contains the key of the nearest parent for the element.

ClickHouse supports the hierarchical property for external dictionary attributes. This property allows you to configure the hierarchical dictionary similar to described above.

The dictGetHierarchy function allows you to get the parent chain of an element.

For our example, the structure of the dictionary can be the following:

DDL

CREATE DICTIONARY regions_dict
(
    region_id UInt64,
    parent_region UInt64 DEFAULT 0 HIERARCHICAL,
    region_name String DEFAULT ''
)
PRIMARY KEY region_id
SOURCE(...)
LAYOUT(HASHED())
LIFETIME(3600);

Configuration file

<dictionary>
    <structure>
        <id>
            <name>region_id</name>
        </id>

        <attribute>
            <name>parent_region</name>
            <type>UInt64</type>
            <null_value>0</null_value>
            <hierarchical>true</hierarchical>
        </attribute>

        <attribute>
            <name>region_name</name>
            <type>String</type>
            <null_value></null_value>
        </attribute>

    </structure>
</dictionary>