ClickHouse/docs/en/sql-reference/dictionaries/index.md

slug	sidebar_label	sidebar_position
/en/sql-reference/dictionaries	Defining Dictionaries	35

import SelfManaged from '@site/docs/en/_snippets/_self_managed_only_no_roadmap.md'; import CloudDetails from '@site/docs/en/sql-reference/dictionaries/_snippet_dictionary_in_cloud.md';

Dictionaries

A dictionary is a mapping (key -> attributes) that is convenient for various types of reference lists.

ClickHouse supports special functions for working with dictionaries that can be used in queries. It is easier and more efficient to use dictionaries with functions than a JOIN with reference tables.

ClickHouse supports:

• Dictionaries with a set of functions.
• Embedded dictionaries with a specific set of functions.

:::tip Tutorial If you are getting started with Dictionaries in ClickHouse we have a tutorial that covers that topic. Take a look here. :::

You can add your own dictionaries from various data sources. The source for a dictionary can be a ClickHouse table, a local text or executable file, an HTTP(s) resource, or another DBMS. For more information, see “Dictionary Sources”.

ClickHouse:

• Fully or partially stores dictionaries in RAM.
• Periodically updates dictionaries and dynamically loads missing values. In other words, dictionaries can be loaded dynamically.
• Allows creating dictionaries with xml files or DDL queries.

The configuration of dictionaries can be located in one or more xml-files. The path to the configuration is specified in the dictionaries_config parameter.

Dictionaries can be loaded at server startup or at first use, depending on the dictionaries_lazy_load setting.

The dictionaries system table contains information about dictionaries configured at server. For each dictionary you can find there:

• Status of the dictionary.
• Configuration parameters.
• Metrics like amount of RAM allocated for the dictionary or a number of queries since the dictionary was successfully loaded.

Creating a dictionary with a DDL query

Dictionaries can be created with DDL queries, and this is the recommended method because with DDL created dictionaries:

• No additional records are added to server configuration files
• The dictionaries can be worked with as first-class entities, like tables or views
• Data can be read directly, using familiar SELECT rather than dictionary table functions
• The dictionaries can be easily renamed

Creating a dictionary with a configuration file

:::note Creating a dictionary with a configuration file is not applicable to ClickHouse Cloud. Please use DDL (see above), and create your dictionary as user default. :::

The dictionary configuration file has the following format:

<clickhouse>
    <comment>An optional element with any content. Ignored by the ClickHouse server.</comment>

    <!--Optional element. File name with substitutions-->
    <include_from>/etc/metrika.xml</include_from>


    <dictionary>
        <!-- Dictionary configuration. -->
        <!-- There can be any number of dictionary sections in a configuration file. -->
    </dictionary>

</clickhouse>

You can configure any number of dictionaries in the same file.

:::note You can convert values for a small dictionary by describing it in a SELECT query (see the transform function). This functionality is not related to dictionaries. :::

Configuring a Dictionary

If dictionary is configured using xml file, than dictionary configuration has the following structure:

<dictionary>
    <name>dict_name</name>

    <structure>
      <!-- Complex key configuration -->
    </structure>

    <source>
      <!-- Source configuration -->
    </source>

    <layout>
      <!-- Memory layout configuration -->
    </layout>

    <lifetime>
      <!-- Lifetime of dictionary in memory -->
    </lifetime>
</dictionary>

Corresponding DDL-query has the following structure:

CREATE DICTIONARY dict_name
(
    ... -- attributes
)
PRIMARY KEY ... -- complex or single key configuration
SOURCE(...) -- Source configuration
LAYOUT(...) -- Memory layout configuration
LIFETIME(...) -- Lifetime of dictionary in memory

Storing Dictionaries in Memory

There are a variety of ways to store dictionaries in memory.

We recommend flat, hashed and complex_key_hashed, which provide optimal processing speed.

Caching is not recommended because of potentially poor performance and difficulties in selecting optimal parameters. Read more in the section cache.

There are several ways to improve dictionary performance:

• Call the function for working with the dictionary after GROUP BY.
• Mark attributes to extract as injective. An attribute is called injective if different attribute values correspond to different keys. So when GROUP BY uses a function that fetches an attribute value by the key, this function is automatically taken out of GROUP BY.

ClickHouse generates an exception for errors with dictionaries. Examples of errors:

• The dictionary being accessed could not be loaded.
• Error querying a cached dictionary.

You can view the list of dictionaries and their statuses in the system.dictionaries table.

The configuration looks like this:

<clickhouse>
    <dictionary>
        ...
        <layout>
            <layout_type>
                <!-- layout settings -->
            </layout_type>
        </layout>
        ...
    </dictionary>
</clickhouse>

Corresponding DDL-query:

CREATE DICTIONARY (...)
...
LAYOUT(LAYOUT_TYPE(param value)) -- layout settings
...

Dictionaries without word complex-key* in a layout have a key with UInt64 type, complex-key* dictionaries have a composite key (complex, with arbitrary types).

UInt64 keys in XML dictionaries are defined with <id> tag.

Configuration example (column key_column has UInt64 type):

...
<structure>
    <id>
        <name>key_column</name>
    </id>
...

Composite complex keys XML dictionaries are defined <key> tag.

Configuration example of a composite key (key has one element with String type):

...
<structure>
    <key>
        <attribute>
            <name>country_code</name>
            <type>String</type>
        </attribute>
    </key>
...

Ways to Store Dictionaries in Memory

• flat
• hashed
• sparse_hashed
• complex_key_hashed
• complex_key_sparse_hashed
• hashed_array
• complex_key_hashed_array
• range_hashed
• complex_key_range_hashed
• cache
• complex_key_cache
• ssd_cache
• complex_key_ssd_cache
• direct
• complex_key_direct
• ip_trie

flat

The dictionary is completely stored in memory in the form of flat arrays. How much memory does the dictionary use? The amount is proportional to the size of the largest key (in space used).

The dictionary key has the UInt64 type and the value is limited to max_array_size (by default — 500,000). If a larger key is discovered when creating the dictionary, ClickHouse throws an exception and does not create the dictionary. Dictionary flat arrays initial size is controlled by initial_array_size setting (by default — 1024).

All types of sources are supported. When updating, data (from a file or from a table) is read in it entirety.

This method provides the best performance among all available methods of storing the dictionary.

Configuration example:

<layout>
  <flat>
    <initial_array_size>50000</initial_array_size>
    <max_array_size>5000000</max_array_size>
  </flat>
</layout>

or

LAYOUT(FLAT(INITIAL_ARRAY_SIZE 50000 MAX_ARRAY_SIZE 5000000))

hashed

The dictionary is completely stored in memory in the form of a hash table. The dictionary can contain any number of elements with any identifiers. In practice, the number of keys can reach tens of millions of items.

The dictionary key has the UInt64 type.

All types of sources are supported. When updating, data (from a file or from a table) is read in its entirety.

Configuration example:

<layout>
  <hashed />
</layout>

or

LAYOUT(HASHED())

Configuration example:

<layout>
  <hashed>
    <!-- If shards greater then 1 (default is `1`) the dictionary will load
         data in parallel, useful if you have huge amount of elements in one
         dictionary. -->
    <shards>10</shards>

    <!-- Size of the backlog for blocks in parallel queue.

         Since the bottleneck in parallel loading is rehash, and so to avoid
         stalling because of thread is doing rehash, you need to have some
         backlog.

         10000 is good balance between memory and speed.
         Even for 10e10 elements and can handle all the load without starvation. -->
    <shard_load_queue_backlog>10000</shard_load_queue_backlog>

    <!-- Maximum load factor of the hash table, with greater values, the memory
         is utilized more efficiently (less memory is wasted) but read/performance
         may deteriorate.

         Valid values: [0.5, 0.99]
         Default: 0.5 -->
    <max_load_factor>0.5</max_load_factor>
  </hashed>
</layout>

or

LAYOUT(HASHED([SHARDS 1] [SHARD_LOAD_QUEUE_BACKLOG 10000] [MAX_LOAD_FACTOR 0.5]))

sparse_hashed

Similar to hashed, but uses less memory in favor more CPU usage.

The dictionary key has the UInt64 type.

Configuration example:

<layout>
  <sparse_hashed>
    <!-- <shards>1</shards> -->
    <!-- <shard_load_queue_backlog>10000</shard_load_queue_backlog> -->
    <!-- <max_load_factor>0.5</max_load_factor> -->
  </sparse_hashed>
</layout>

or

LAYOUT(SPARSE_HASHED([SHARDS 1] [SHARD_LOAD_QUEUE_BACKLOG 10000] [MAX_LOAD_FACTOR 0.5]))

It is also possible to use shards for this type of dictionary, and again it is more important for sparse_hashed then for hashed, since sparse_hashed is slower.

complex_key_hashed

This type of storage is for use with composite keys. Similar to hashed.

Configuration example:

<layout>
  <complex_key_hashed>
    <!-- <shards>1</shards> -->
    <!-- <shard_load_queue_backlog>10000</shard_load_queue_backlog> -->
    <!-- <max_load_factor>0.5</max_load_factor> -->
  </complex_key_hashed>
</layout>

or

LAYOUT(COMPLEX_KEY_HASHED([SHARDS 1] [SHARD_LOAD_QUEUE_BACKLOG 10000] [MAX_LOAD_FACTOR 0.5]))

complex_key_sparse_hashed

This type of storage is for use with composite keys. Similar to sparse_hashed.

Configuration example:

<layout>
  <complex_key_sparse_hashed>
    <!-- <shards>1</shards> -->
    <!-- <shard_load_queue_backlog>10000</shard_load_queue_backlog> -->
    <!-- <max_load_factor>0.5</max_load_factor> -->
  </complex_key_sparse_hashed>
</layout>

or

LAYOUT(COMPLEX_KEY_SPARSE_HASHED([SHARDS 1] [SHARD_LOAD_QUEUE_BACKLOG 10000] [MAX_LOAD_FACTOR 0.5]))

hashed_array

The dictionary is completely stored in memory. Each attribute is stored in an array. The key attribute is stored in the form of a hashed table where value is an index in the attributes array. The dictionary can contain any number of elements with any identifiers. In practice, the number of keys can reach tens of millions of items.

The dictionary key has the UInt64 type.

All types of sources are supported. When updating, data (from a file or from a table) is read in its entirety.

Configuration example:

<layout>
  <hashed_array>
  </hashed_array>
</layout>

or

LAYOUT(HASHED_ARRAY([SHARDS 1]))

complex_key_hashed_array

This type of storage is for use with composite keys. Similar to hashed_array.

Configuration example:

<layout>
  <complex_key_hashed_array />
</layout>

or

LAYOUT(COMPLEX_KEY_HASHED_ARRAY([SHARDS 1]))

range_hashed

The dictionary is stored in memory in the form of a hash table with an ordered array of ranges and their corresponding values.

The dictionary key has the UInt64 type. This storage method works the same way as hashed and allows using date/time (arbitrary numeric type) ranges in addition to the key.

Example: The table contains discounts for each advertiser in the format:

┌─advertiser_id─┬─discount_start_date─┬─discount_end_date─┬─amount─┐
│           123 │          2015-01-16 │        2015-01-31 │   0.25 │
│           123 │          2015-01-01 │        2015-01-15 │   0.15 │
│           456 │          2015-01-01 │        2015-01-15 │   0.05 │
└───────────────┴─────────────────────┴───────────────────┴────────┘

To use a sample for date ranges, define the range_min and range_max elements in the structure. These elements must contain elements name and type (if type is not specified, the default type will be used - Date). type can be any numeric type (Date / DateTime / UInt64 / Int32 / others).

:::note Values of range_min and range_max should fit in Int64 type. :::

Example:

<layout>
    <range_hashed>
        <!-- Strategy for overlapping ranges (min/max). Default: min (return a matching range with the min(range_min -> range_max) value) -->
        <range_lookup_strategy>min</range_lookup_strategy>
    </range_hashed>
</layout>
<structure>
    <id>
        <name>advertiser_id</name>
    </id>
    <range_min>
        <name>discount_start_date</name>
        <type>Date</type>
    </range_min>
    <range_max>
        <name>discount_end_date</name>
        <type>Date</type>
    </range_max>
    ...

or

CREATE DICTIONARY discounts_dict (
    advertiser_id UInt64,
    discount_start_date Date,
    discount_end_date Date,
    amount Float64
)
PRIMARY KEY id
SOURCE(CLICKHOUSE(TABLE 'discounts'))
LIFETIME(MIN 1 MAX 1000)
LAYOUT(RANGE_HASHED(range_lookup_strategy 'max'))
RANGE(MIN discount_start_date MAX discount_end_date)

To work with these dictionaries, you need to pass an additional argument to the dictGet function, for which a range is selected:

dictGet('dict_name', 'attr_name', id, date)

Query example:

SELECT dictGet('discounts_dict', 'amount', 1, '2022-10-20'::Date);

This function returns the value for the specified ids and the date range that includes the passed date.

Details of the algorithm:

• If the id is not found or a range is not found for the id, it returns the default value of the attribute's type.
• If there are overlapping ranges and range_lookup_strategy=min, it returns a matching range with minimal range_min, if several ranges found, it returns a range with minimal range_max, if again several ranges found (several ranges had the same range_min and range_max it returns a random range of them.
• If there are overlapping ranges and range_lookup_strategy=max, it returns a matching range with maximal range_min, if several ranges found, it returns a range with maximal range_max, if again several ranges found (several ranges had the same range_min and range_max it returns a random range of them.
• If the range_max is NULL, the range is open. NULL is treated as maximal possible value. For the range_min 1970-01-01 or 0 (-MAX_INT) can be used as the open value.

Configuration example:

<clickhouse>
    <dictionary>
        ...

        <layout>
            <range_hashed />
        </layout>

        <structure>
            <id>
                <name>Abcdef</name>
            </id>
            <range_min>
                <name>StartTimeStamp</name>
                <type>UInt64</type>
            </range_min>
            <range_max>
                <name>EndTimeStamp</name>
                <type>UInt64</type>
            </range_max>
            <attribute>
                <name>XXXType</name>
                <type>String</type>
                <null_value />
            </attribute>
        </structure>

    </dictionary>
</clickhouse>

or

CREATE DICTIONARY somedict(
    Abcdef UInt64,
    StartTimeStamp UInt64,
    EndTimeStamp UInt64,
    XXXType String DEFAULT ''
)
PRIMARY KEY Abcdef
RANGE(MIN StartTimeStamp MAX EndTimeStamp)

Configuration example with overlapping ranges and open ranges:

CREATE TABLE discounts
(
    advertiser_id UInt64,
    discount_start_date Date,
    discount_end_date Nullable(Date),
    amount Float64
)
ENGINE = Memory;

INSERT INTO discounts VALUES (1, '2015-01-01', Null, 0.1);
INSERT INTO discounts VALUES (1, '2015-01-15', Null, 0.2);
INSERT INTO discounts VALUES (2, '2015-01-01', '2015-01-15', 0.3);
INSERT INTO discounts VALUES (2, '2015-01-04', '2015-01-10', 0.4);
INSERT INTO discounts VALUES (3, '1970-01-01', '2015-01-15', 0.5);
INSERT INTO discounts VALUES (3, '1970-01-01', '2015-01-10', 0.6);

SELECT * FROM discounts ORDER BY advertiser_id, discount_start_date;
┌─advertiser_id─┬─discount_start_date─┬─discount_end_date─┬─amount─┐
│             1 │          2015-01-01 │              ᴺᵁᴸᴸ │    0.1 │
│             1 │          2015-01-15 │              ᴺᵁᴸᴸ │    0.2 │
│             2 │          2015-01-01 │        2015-01-15 │    0.3 │
│             2 │          2015-01-04 │        2015-01-10 │    0.4 │
│             3 │          1970-01-01 │        2015-01-15 │    0.5 │
│             3 │          1970-01-01 │        2015-01-10 │    0.6 │
└───────────────┴─────────────────────┴───────────────────┴────────┘

-- RANGE_LOOKUP_STRATEGY 'max'

CREATE DICTIONARY discounts_dict
(
    advertiser_id UInt64,
    discount_start_date Date,
    discount_end_date Nullable(Date),
    amount Float64
)
PRIMARY KEY advertiser_id
SOURCE(CLICKHOUSE(TABLE discounts))
LIFETIME(MIN 600 MAX 900)
LAYOUT(RANGE_HASHED(RANGE_LOOKUP_STRATEGY 'max'))
RANGE(MIN discount_start_date MAX discount_end_date);

select dictGet('discounts_dict', 'amount', 1, toDate('2015-01-14')) res;
┌─res─┐
│ 0.1 │ -- the only one range is matching: 2015-01-01 - Null
└─────┘

select dictGet('discounts_dict', 'amount', 1, toDate('2015-01-16')) res;
┌─res─┐
│ 0.2 │ -- two ranges are matching, range_min 2015-01-15 (0.2) is bigger than 2015-01-01 (0.1)
└─────┘

select dictGet('discounts_dict', 'amount', 2, toDate('2015-01-06')) res;
┌─res─┐
│ 0.4 │ -- two ranges are matching, range_min 2015-01-04 (0.4) is bigger than 2015-01-01 (0.3)
└─────┘

select dictGet('discounts_dict', 'amount', 3, toDate('2015-01-01')) res;
┌─res─┐
│ 0.5 │ -- two ranges are matching, range_min are equal, 2015-01-15 (0.5) is bigger than 2015-01-10 (0.6)
└─────┘

DROP DICTIONARY discounts_dict;

-- RANGE_LOOKUP_STRATEGY 'min'

CREATE DICTIONARY discounts_dict
(
    advertiser_id UInt64,
    discount_start_date Date,
    discount_end_date Nullable(Date),
    amount Float64
)
PRIMARY KEY advertiser_id
SOURCE(CLICKHOUSE(TABLE discounts))
LIFETIME(MIN 600 MAX 900)
LAYOUT(RANGE_HASHED(RANGE_LOOKUP_STRATEGY 'min'))
RANGE(MIN discount_start_date MAX discount_end_date);

select dictGet('discounts_dict', 'amount', 1, toDate('2015-01-14')) res;
┌─res─┐
│ 0.1 │ -- the only one range is matching: 2015-01-01 - Null
└─────┘

select dictGet('discounts_dict', 'amount', 1, toDate('2015-01-16')) res;
┌─res─┐
│ 0.1 │ -- two ranges are matching, range_min 2015-01-01 (0.1) is less than 2015-01-15 (0.2)
└─────┘

select dictGet('discounts_dict', 'amount', 2, toDate('2015-01-06')) res;
┌─res─┐
│ 0.3 │ -- two ranges are matching, range_min 2015-01-01 (0.3) is less than 2015-01-04 (0.4)
└─────┘

select dictGet('discounts_dict', 'amount', 3, toDate('2015-01-01')) res;
┌─res─┐
│ 0.6 │ -- two ranges are matching, range_min are equal, 2015-01-10 (0.6) is less than 2015-01-15 (0.5)
└─────┘

complex_key_range_hashed

The dictionary is stored in memory in the form of a hash table with an ordered array of ranges and their corresponding values (see range_hashed). This type of storage is for use with composite keys.

Configuration example:

CREATE DICTIONARY range_dictionary
(
  CountryID UInt64,
  CountryKey String,
  StartDate Date,
  EndDate Date,
  Tax Float64 DEFAULT 0.2
)
PRIMARY KEY CountryID, CountryKey
SOURCE(CLICKHOUSE(TABLE 'date_table'))
LIFETIME(MIN 1 MAX 1000)
LAYOUT(COMPLEX_KEY_RANGE_HASHED())
RANGE(MIN StartDate MAX EndDate);

cache

The dictionary is stored in a cache that has a fixed number of cells. These cells contain frequently used elements.

The dictionary key has the UInt64 type.

When searching for a dictionary, the cache is searched first. For each block of data, all keys that are not found in the cache or are outdated are requested from the source using SELECT attrs... FROM db.table WHERE id IN (k1, k2, ...). The received data is then written to the cache.

If keys are not found in dictionary, then update cache task is created and added into update queue. Update queue properties can be controlled with settings max_update_queue_size, update_queue_push_timeout_milliseconds, query_wait_timeout_milliseconds, max_threads_for_updates.

For cache dictionaries, the expiration lifetime of data in the cache can be set. If more time than lifetime has passed since loading the data in a cell, the cell’s value is not used and key becomes expired. The key is re-requested the next time it needs to be used. This behaviour can be configured with setting allow_read_expired_keys.

This is the least effective of all the ways to store dictionaries. The speed of the cache depends strongly on correct settings and the usage scenario. A cache type dictionary performs well only when the hit rates are high enough (recommended 99% and higher). You can view the average hit rate in the system.dictionaries table.

If setting allow_read_expired_keys is set to 1, by default 0. Then dictionary can support asynchronous updates. If a client requests keys and all of them are in cache, but some of them are expired, then dictionary will return expired keys for a client and request them asynchronously from the source.

To improve cache performance, use a subquery with LIMIT, and call the function with the dictionary externally.

All types of sources are supported.

Example of settings:

<layout>
    <cache>
        <!-- The size of the cache, in number of cells. Rounded up to a power of two. -->
        <size_in_cells>1000000000</size_in_cells>
        <!-- Allows to read expired keys. -->
        <allow_read_expired_keys>0</allow_read_expired_keys>
        <!-- Max size of update queue. -->
        <max_update_queue_size>100000</max_update_queue_size>
        <!-- Max timeout in milliseconds for push update task into queue. -->
        <update_queue_push_timeout_milliseconds>10</update_queue_push_timeout_milliseconds>
        <!-- Max wait timeout in milliseconds for update task to complete. -->
        <query_wait_timeout_milliseconds>60000</query_wait_timeout_milliseconds>
        <!-- Max threads for cache dictionary update. -->
        <max_threads_for_updates>4</max_threads_for_updates>
    </cache>
</layout>

or

LAYOUT(CACHE(SIZE_IN_CELLS 1000000000))

Set a large enough cache size. You need to experiment to select the number of cells:

1. Set some value.
2. Run queries until the cache is completely full.
3. Assess memory consumption using the system.dictionaries table.
4. Increase or decrease the number of cells until the required memory consumption is reached.

:::note Do not use ClickHouse as a source, because it is slow to process queries with random reads. :::

complex_key_cache

This type of storage is for use with composite keys. Similar to cache.

ssd_cache

Similar to cache, but stores data on SSD and index in RAM. All cache dictionary settings related to update queue can also be applied to SSD cache dictionaries.

The dictionary key has the UInt64 type.

<layout>
    <ssd_cache>
        <!-- Size of elementary read block in bytes. Recommended to be equal to SSD's page size. -->
        <block_size>4096</block_size>
        <!-- Max cache file size in bytes. -->
        <file_size>16777216</file_size>
        <!-- Size of RAM buffer in bytes for reading elements from SSD. -->
        <read_buffer_size>131072</read_buffer_size>
        <!-- Size of RAM buffer in bytes for aggregating elements before flushing to SSD. -->
        <write_buffer_size>1048576</write_buffer_size>
        <!-- Path where cache file will be stored. -->
        <path>/var/lib/clickhouse/user_files/test_dict</path>
    </ssd_cache>
</layout>

or

LAYOUT(SSD_CACHE(BLOCK_SIZE 4096 FILE_SIZE 16777216 READ_BUFFER_SIZE 1048576
    PATH '/var/lib/clickhouse/user_files/test_dict'))

complex_key_ssd_cache

This type of storage is for use with composite keys. Similar to ssd_cache.

direct

The dictionary is not stored in memory and directly goes to the source during the processing of a request.

The dictionary key has the UInt64 type.

All types of sources, except local files, are supported.

Configuration example:

<layout>
  <direct />
</layout>

or

LAYOUT(DIRECT())

complex_key_direct

This type of storage is for use with composite keys. Similar to direct.

ip_trie

This type of storage is for mapping network prefixes (IP addresses) to metadata such as ASN.

Example

Suppose we have a table in ClickHouse that contains our IP prefixes and mappings:

CREATE TABLE my_ip_addresses (
	prefix String,
	asn UInt32,
	cca2 String
)
ENGINE = MergeTree
PRIMARY KEY prefix;

INSERT INTO my_ip_addresses VALUES
	('202.79.32.0/20', 17501, 'NP'),
    ('2620:0:870::/48', 3856, 'US'),
    ('2a02:6b8:1::/48', 13238, 'RU'),
    ('2001:db8::/32', 65536, 'ZZ')
;

Let's define an ip_trie dictionary for this table. The ip_trie layout requires a composite key:

<structure>
    <key>
        <attribute>
            <name>prefix</name>
            <type>String</type>
        </attribute>
    </key>
    <attribute>
            <name>asn</name>
            <type>UInt32</type>
            <null_value />
    </attribute>
    <attribute>
            <name>cca2</name>
            <type>String</type>
            <null_value>??</null_value>
    </attribute>
    ...
</structure>
<layout>
    <ip_trie>
        <!-- Key attribute `prefix` can be retrieved via dictGetString. -->
        <!-- This option increases memory usage. -->
        <access_to_key_from_attributes>true</access_to_key_from_attributes>
    </ip_trie>
</layout>

or

CREATE DICTIONARY my_ip_trie_dictionary (
    prefix String,
    asn UInt32,
    cca2 String DEFAULT '??'
)
PRIMARY KEY prefix
SOURCE(CLICKHOUSE(TABLE 'my_ip_addresses'))
LAYOUT(IP_TRIE)
LIFETIME(3600);

The key must have only one String type attribute that contains an allowed IP prefix. Other types are not supported yet.

The syntax is:

dictGetT('dict_name', 'attr_name', ip)

The function takes either UInt32 for IPv4, or FixedString(16) for IPv6. For example:

SELECT dictGet('my_ip_trie_dictionary', 'cca2', toIPv4('202.79.32.10')) AS result;

┌─result─┐
│ NP     │
└────────┘


SELECT dictGet('my_ip_trie_dictionary', 'asn', IPv6StringToNum('2001:db8::1')) AS result;

┌─result─┐
│  65536 │
└────────┘


SELECT dictGet('my_ip_trie_dictionary', ('asn', 'cca2'), IPv6StringToNum('2001:db8::1')) AS result;

┌─result───────┐
│ (65536,'ZZ') │
└──────────────┘

Other types are not supported yet. The function returns the attribute for the prefix that corresponds to this IP address. If there are overlapping prefixes, the most specific one is returned.

Data must completely fit into RAM.

Refreshing dictionary data using LIFETIME

ClickHouse periodically updates dictionaries based on the LIFETIME tag (defined in seconds). LIFETIME is the update interval for fully downloaded dictionaries and the invalidation interval for cached dictionaries.

During updates, the old version of a dictionary can still be queried. Dictionary updates (other than when loading the dictionary for first use) do not block queries. If an error occurs during an update, the error is written to the server log and queries can continue using the old version of the dictionary. If a dictionary update is successful, the old version of the dictionary is replaced atomically.

Example of settings:

<dictionary>
    ...
    <lifetime>300</lifetime>
    ...
</dictionary>

or

CREATE DICTIONARY (...)
...
LIFETIME(300)
...

Setting <lifetime>0</lifetime> (LIFETIME(0)) prevents dictionaries from updating.

You can set a time interval for updates, and ClickHouse will choose a uniformly random time within this range. This is necessary in order to distribute the load on the dictionary source when updating on a large number of servers.

Example of settings:

<dictionary>
    ...
    <lifetime>
        <min>300</min>
        <max>360</max>
    </lifetime>
    ...
</dictionary>

or

LIFETIME(MIN 300 MAX 360)

If <min>0</min> and <max>0</max>, ClickHouse does not reload the dictionary by timeout. In this case, ClickHouse can reload the dictionary earlier if the dictionary configuration file was changed or the SYSTEM RELOAD DICTIONARY command was executed.

When updating the dictionaries, the ClickHouse server applies different logic depending on the type of source:

• For a text file, it checks the time of modification. If the time differs from the previously recorded time, the dictionary is updated.
• For MySQL source, the time of modification is checked using a SHOW TABLE STATUS query (in case of MySQL 8 you need to disable meta-information caching in MySQL by set global information_schema_stats_expiry=0).
• Dictionaries from other sources are updated every time by default.

For other sources (ODBC, PostgreSQL, ClickHouse, etc), you can set up a query that will update the dictionaries only if they really changed, rather than each time. To do this, follow these steps:

• The dictionary table must have a field that always changes when the source data is updated.
• The settings of the source must specify a query that retrieves the changing field. The ClickHouse server interprets the query result as a row, and if this row has changed relative to its previous state, the dictionary is updated. Specify the query in the <invalidate_query> field in the settings for the source.

Example of settings:

<dictionary>
    ...
    <odbc>
      ...
      <invalidate_query>SELECT update_time FROM dictionary_source where id = 1</invalidate_query>
    </odbc>
    ...
</dictionary>

or

...
SOURCE(ODBC(... invalidate_query 'SELECT update_time FROM dictionary_source where id = 1'))
...

For Cache, ComplexKeyCache, SSDCache, and SSDComplexKeyCache dictionaries both synchronous and asynchronous updates are supported.

It is also possible for Flat, Hashed, ComplexKeyHashed dictionaries to only request data that was changed after the previous update. If update_field is specified as part of the dictionary source configuration, value of the previous update time in seconds will be added to the data request. Depends on source type (Executable, HTTP, MySQL, PostgreSQL, ClickHouse, or ODBC) different logic will be applied to update_field before request data from an external source.

• If the source is HTTP then update_field will be added as a query parameter with the last update time as the parameter value.
• If the source is Executable then update_field will be added as an executable script argument with the last update time as the argument value.
• If the source is ClickHouse, MySQL, PostgreSQL, ODBC there will be an additional part of WHERE, where update_field is compared as greater or equal with the last update time.
  • Per default, this WHERE-condition is checked at the highest level of the SQL-Query. Alternatively, the condition can be checked in any other WHERE-clause within the query using the {condition}-keyword. Example:

  ...
  SOURCE(CLICKHOUSE(...
      update_field 'added_time'
      QUERY '
          SELECT my_arr.1 AS x, my_arr.2 AS y, creation_time
          FROM (
              SELECT arrayZip(x_arr, y_arr) AS my_arr, creation_time
              FROM dictionary_source
              WHERE {condition}
          )'
  ))
  ...
  

If update_field option is set, additional option update_lag can be set. Value of update_lag option is subtracted from previous update time before request updated data.

Example of settings:

<dictionary>
    ...
        <clickhouse>
            ...
            <update_field>added_time</update_field>
            <update_lag>15</update_lag>
        </clickhouse>
    ...
</dictionary>

or

...
SOURCE(CLICKHOUSE(... update_field 'added_time' update_lag 15))
...

Dictionary Sources

A dictionary can be connected to ClickHouse from many different sources.

If the dictionary is configured using an xml-file, the configuration looks like this:

<clickhouse>
  <dictionary>
    ...
    <source>
      <source_type>
        <!-- Source configuration -->
      </source_type>
    </source>
    ...
  </dictionary>
  ...
</clickhouse>

In case of DDL-query, the configuration described above will look like:

CREATE DICTIONARY dict_name (...)
...
SOURCE(SOURCE_TYPE(param1 val1 ... paramN valN)) -- Source configuration
...

The source is configured in the source section.

For source types Local file, Executable file, HTTP(s), ClickHouse optional settings are available:

<source>
  <file>
    <path>/opt/dictionaries/os.tsv</path>
    <format>TabSeparated</format>
  </file>
  <settings>
      <format_csv_allow_single_quotes>0</format_csv_allow_single_quotes>
  </settings>
</source>

or

SOURCE(FILE(path './user_files/os.tsv' format 'TabSeparated'))
SETTINGS(format_csv_allow_single_quotes = 0)

Types of sources (source_type):

• Local file
• Executable File
• Executable Pool
• HTTP(S)
• DBMS
  • ODBC
  • MySQL
  • ClickHouse
  • MongoDB
  • Redis
  • Cassandra
  • PostgreSQL

Local File

Example of settings:

<source>
  <file>
    <path>/opt/dictionaries/os.tsv</path>
    <format>TabSeparated</format>
  </file>
</source>

or

SOURCE(FILE(path './user_files/os.tsv' format 'TabSeparated'))

Setting fields:

• path – The absolute path to the file.
• format – The file format. All the formats described in Formats are supported.

When a dictionary with source FILE is created via DDL command (CREATE DICTIONARY ...), the source file needs to be located in the user_files directory to prevent DB users from accessing arbitrary files on the ClickHouse node.

See Also

• Dictionary function

Executable File

Working with executable files depends on how the dictionary is stored in memory. If the dictionary is stored using cache and complex_key_cache, ClickHouse requests the necessary keys by sending a request to the executable file’s STDIN. Otherwise, ClickHouse starts the executable file and treats its output as dictionary data.

Example of settings:

<source>
    <executable>
        <command>cat /opt/dictionaries/os.tsv</command>
        <format>TabSeparated</format>
        <implicit_key>false</implicit_key>
    </executable>
</source>

Setting fields:

• command — The absolute path to the executable file, or the file name (if the command's directory is in the PATH).
• format — The file format. All the formats described in Formats are supported.
• command_termination_timeout — The executable script should contain a main read-write loop. After the dictionary is destroyed, the pipe is closed, and the executable file will have command_termination_timeout seconds to shutdown before ClickHouse will send a SIGTERM signal to the child process. command_termination_timeout is specified in seconds. Default value is 10. Optional parameter.
• command_read_timeout - Timeout for reading data from command stdout in milliseconds. Default value 10000. Optional parameter.
• command_write_timeout - Timeout for writing data to command stdin in milliseconds. Default value 10000. Optional parameter.
• implicit_key — The executable source file can return only values, and the correspondence to the requested keys is determined implicitly — by the order of rows in the result. Default value is false.
• execute_direct - If execute_direct = 1, then command will be searched inside user_scripts folder specified by user_scripts_path. Additional script arguments can be specified using a whitespace separator. Example: script_name arg1 arg2. If execute_direct = 0, command is passed as argument for bin/sh -c. Default value is 0. Optional parameter.
• send_chunk_header - controls whether to send row count before sending a chunk of data to process. Optional. Default value is false.

That dictionary source can be configured only via XML configuration. Creating dictionaries with executable source via DDL is disabled; otherwise, the DB user would be able to execute arbitrary binaries on the ClickHouse node.

Executable Pool

Executable pool allows loading data from pool of processes. This source does not work with dictionary layouts that need to load all data from source. Executable pool works if the dictionary is stored using cache, complex_key_cache, ssd_cache, complex_key_ssd_cache, direct, or complex_key_direct layouts.

Executable pool will spawn a pool of processes with the specified command and keep them running until they exit. The program should read data from STDIN while it is available and output the result to STDOUT. It can wait for the next block of data on STDIN. ClickHouse will not close STDIN after processing a block of data, but will pipe another chunk of data when needed. The executable script should be ready for this way of data processing — it should poll STDIN and flush data to STDOUT early.

Example of settings:

<source>
    <executable_pool>
        <command><command>while read key; do printf "$key\tData for key $key\n"; done</command</command>
        <format>TabSeparated</format>
        <pool_size>10</pool_size>
        <max_command_execution_time>10<max_command_execution_time>
        <implicit_key>false</implicit_key>
    </executable_pool>
</source>

Setting fields:

• command — The absolute path to the executable file, or the file name (if the program directory is written to PATH).
• format — The file format. All the formats described in “Formats” are supported.
• pool_size — Size of pool. If 0 is specified as pool_size then there is no pool size restrictions. Default value is 16.
• command_termination_timeout — executable script should contain main read-write loop. After dictionary is destroyed, pipe is closed, and executable file will have command_termination_timeout seconds to shutdown, before ClickHouse will send SIGTERM signal to child process. Specified in seconds. Default value is 10. Optional parameter.
• max_command_execution_time — Maximum executable script command execution time for processing block of data. Specified in seconds. Default value is 10. Optional parameter.
• command_read_timeout - timeout for reading data from command stdout in milliseconds. Default value 10000. Optional parameter.
• command_write_timeout - timeout for writing data to command stdin in milliseconds. Default value 10000. Optional parameter.
• implicit_key — The executable source file can return only values, and the correspondence to the requested keys is determined implicitly — by the order of rows in the result. Default value is false. Optional parameter.
• execute_direct - If execute_direct = 1, then command will be searched inside user_scripts folder specified by user_scripts_path. Additional script arguments can be specified using whitespace separator. Example: script_name arg1 arg2. If execute_direct = 0, command is passed as argument for bin/sh -c. Default value is 1. Optional parameter.
• send_chunk_header - controls whether to send row count before sending a chunk of data to process. Optional. Default value is false.

That dictionary source can be configured only via XML configuration. Creating dictionaries with executable source via DDL is disabled, otherwise, the DB user would be able to execute arbitrary binary on ClickHouse node.

HTTP(S)

Working with an HTTP(S) server depends on how the dictionary is stored in memory. If the dictionary is stored using cache and complex_key_cache, ClickHouse requests the necessary keys by sending a request via the POST method.

Example of settings:

<source>
    <http>
        <url>http://[::1]/os.tsv</url>
        <format>TabSeparated</format>
        <credentials>
            <user>user</user>
            <password>password</password>
        </credentials>
        <headers>
            <header>
                <name>API-KEY</name>
                <value>key</value>
            </header>
        </headers>
    </http>
</source>

or

SOURCE(HTTP(
    url 'http://[::1]/os.tsv'
    format 'TabSeparated'
    credentials(user 'user' password 'password')
    headers(header(name 'API-KEY' value 'key'))
))

In order for ClickHouse to access an HTTPS resource, you must configure openSSL in the server configuration.

Setting fields:

• url – The source URL.
• format – The file format. All the formats described in “Formats” are supported.
• credentials – Basic HTTP authentication. Optional parameter.
• user – Username required for the authentication.
• password – Password required for the authentication.
• headers – All custom HTTP headers entries used for the HTTP request. Optional parameter.
• header – Single HTTP header entry.
• name – Identifier name used for the header send on the request.
• value – Value set for a specific identifier name.

When creating a dictionary using the DDL command (CREATE DICTIONARY ...) remote hosts for HTTP dictionaries are checked against the contents of remote_url_allow_hosts section from config to prevent database users to access arbitrary HTTP server.

DBMS

ODBC

You can use this method to connect any database that has an ODBC driver.

Example of settings:

<source>
    <odbc>
        <db>DatabaseName</db>
        <table>ShemaName.TableName</table>
        <connection_string>DSN=some_parameters</connection_string>
        <invalidate_query>SQL_QUERY</invalidate_query>
        <query>SELECT id, value_1, value_2 FROM ShemaName.TableName</query>
    </odbc>
</source>

or

SOURCE(ODBC(
    db 'DatabaseName'
    table 'SchemaName.TableName'
    connection_string 'DSN=some_parameters'
    invalidate_query 'SQL_QUERY'
    query 'SELECT id, value_1, value_2 FROM db_name.table_name'
))

Setting fields:

• db – Name of the database. Omit it if the database name is set in the <connection_string> parameters.
• table – Name of the table and schema if exists.
• connection_string – Connection string.
• invalidate_query – Query for checking the dictionary status. Optional parameter. Read more in the section Refreshing dictionary data using LIFETIME.
• query – The custom query. Optional parameter.

:::note The table and query fields cannot be used together. And either one of the table or query fields must be declared. :::

ClickHouse receives quoting symbols from ODBC-driver and quote all settings in queries to driver, so it’s necessary to set table name accordingly to table name case in database.

If you have a problems with encodings when using Oracle, see the corresponding FAQ item.

Known Vulnerability of the ODBC Dictionary Functionality

:::note When connecting to the database through the ODBC driver connection parameter Servername can be substituted. In this case values of USERNAME and PASSWORD from odbc.ini are sent to the remote server and can be compromised. :::

Example of insecure use

Let’s configure unixODBC for PostgreSQL. Content of /etc/odbc.ini:

[gregtest]
Driver = /usr/lib/psqlodbca.so
Servername = localhost
PORT = 5432
DATABASE = test_db
#OPTION = 3
USERNAME = test
PASSWORD = test

If you then make a query such as

SELECT * FROM odbc('DSN=gregtest;Servername=some-server.com', 'test_db');

ODBC driver will send values of USERNAME and PASSWORD from odbc.ini to some-server.com.

Example of Connecting Postgresql

Ubuntu OS.

Installing unixODBC and the ODBC driver for PostgreSQL:

$ sudo apt-get install -y unixodbc odbcinst odbc-postgresql

Configuring /etc/odbc.ini (or ~/.odbc.ini if you signed in under a user that runs ClickHouse):

    [DEFAULT]
    Driver = myconnection

    [myconnection]
    Description         = PostgreSQL connection to my_db
    Driver              = PostgreSQL Unicode
    Database            = my_db
    Servername          = 127.0.0.1
    UserName            = username
    Password            = password
    Port                = 5432
    Protocol            = 9.3
    ReadOnly            = No
    RowVersioning       = No
    ShowSystemTables    = No
    ConnSettings        =

The dictionary configuration in ClickHouse:

<clickhouse>
    <dictionary>
        <name>table_name</name>
        <source>
            <odbc>
                <!-- You can specify the following parameters in connection_string: -->
                <!-- DSN=myconnection;UID=username;PWD=password;HOST=127.0.0.1;PORT=5432;DATABASE=my_db -->
                <connection_string>DSN=myconnection</connection_string>
                <table>postgresql_table</table>
            </odbc>
        </source>
        <lifetime>
            <min>300</min>
            <max>360</max>
        </lifetime>
        <layout>
            <hashed/>
        </layout>
        <structure>
            <id>
                <name>id</name>
            </id>
            <attribute>
                <name>some_column</name>
                <type>UInt64</type>
                <null_value>0</null_value>
            </attribute>
        </structure>
    </dictionary>
</clickhouse>

or

CREATE DICTIONARY table_name (
    id UInt64,
    some_column UInt64 DEFAULT 0
)
PRIMARY KEY id
SOURCE(ODBC(connection_string 'DSN=myconnection' table 'postgresql_table'))
LAYOUT(HASHED())
LIFETIME(MIN 300 MAX 360)

You may need to edit odbc.ini to specify the full path to the library with the driver DRIVER=/usr/local/lib/psqlodbcw.so.

Example of Connecting MS SQL Server

Ubuntu OS.

Installing the ODBC driver for connecting to MS SQL:

$ sudo apt-get install tdsodbc freetds-bin sqsh

Configuring the driver:

    $ cat /etc/freetds/freetds.conf
    ...

    [MSSQL]
    host = 192.168.56.101
    port = 1433
    tds version = 7.0
    client charset = UTF-8

    # test TDS connection
    $ sqsh -S MSSQL -D database -U user -P password


    $ cat /etc/odbcinst.ini

    [FreeTDS]
    Description     = FreeTDS
    Driver          = /usr/lib/x86_64-linux-gnu/odbc/libtdsodbc.so
    Setup           = /usr/lib/x86_64-linux-gnu/odbc/libtdsS.so
    FileUsage       = 1
    UsageCount      = 5

    $ cat /etc/odbc.ini
    # $ cat ~/.odbc.ini # if you signed in under a user that runs ClickHouse

    [MSSQL]
    Description     = FreeTDS
    Driver          = FreeTDS
    Servername      = MSSQL
    Database        = test
    UID             = test
    PWD             = test
    Port            = 1433


    # (optional) test ODBC connection (to use isql-tool install the [unixodbc](https://packages.debian.org/sid/unixodbc)-package)
    $ isql -v MSSQL "user" "password"

Remarks:

• to determine the earliest TDS version that is supported by a particular SQL Server version, refer to the product documentation or look at MS-TDS Product Behavior

Configuring the dictionary in ClickHouse:

<clickhouse>
    <dictionary>
        <name>test</name>
        <source>
            <odbc>
                <table>dict</table>
                <connection_string>DSN=MSSQL;UID=test;PWD=test</connection_string>
            </odbc>
        </source>

        <lifetime>
            <min>300</min>
            <max>360</max>
        </lifetime>

        <layout>
            <flat />
        </layout>

        <structure>
            <id>
                <name>k</name>
            </id>
            <attribute>
                <name>s</name>
                <type>String</type>
                <null_value></null_value>
            </attribute>
        </structure>
    </dictionary>
</clickhouse>

or

CREATE DICTIONARY test (
    k UInt64,
    s String DEFAULT ''
)
PRIMARY KEY k
SOURCE(ODBC(table 'dict' connection_string 'DSN=MSSQL;UID=test;PWD=test'))
LAYOUT(FLAT())
LIFETIME(MIN 300 MAX 360)

Mysql

Example of settings:

<source>
  <mysql>
      <port>3306</port>
      <user>clickhouse</user>
      <password>qwerty</password>
      <replica>
          <host>example01-1</host>
          <priority>1</priority>
      </replica>
      <replica>
          <host>example01-2</host>
          <priority>1</priority>
      </replica>
      <db>db_name</db>
      <table>table_name</table>
      <where>id=10</where>
      <invalidate_query>SQL_QUERY</invalidate_query>
      <fail_on_connection_loss>true</fail_on_connection_loss>
      <query>SELECT id, value_1, value_2 FROM db_name.table_name</query>
  </mysql>
</source>

or

SOURCE(MYSQL(
    port 3306
    user 'clickhouse'
    password 'qwerty'
    replica(host 'example01-1' priority 1)
    replica(host 'example01-2' priority 1)
    db 'db_name'
    table 'table_name'
    where 'id=10'
    invalidate_query 'SQL_QUERY'
    fail_on_connection_loss 'true'
    query 'SELECT id, value_1, value_2 FROM db_name.table_name'
))

Setting fields:

• port – The port on the MySQL server. You can specify it for all replicas, or for each one individually (inside <replica>).

• user – Name of the MySQL user. You can specify it for all replicas, or for each one individually (inside <replica>).

• password – Password of the MySQL user. You can specify it for all replicas, or for each one individually (inside <replica>).

• replica – Section of replica configurations. There can be multiple sections.

    - `replica/host` – The MySQL host.
    - `replica/priority` – The replica priority. When attempting to connect, ClickHouse traverses the replicas in order of priority. The lower the number, the higher the priority.
  

• db – Name of the database.

• table – Name of the table.

• where – The selection criteria. The syntax for conditions is the same as for WHERE clause in MySQL, for example, id > 10 AND id < 20. Optional parameter.

• invalidate_query – Query for checking the dictionary status. Optional parameter. Read more in the section Refreshing dictionary data using LIFETIME.

• fail_on_connection_loss – The configuration parameter that controls behavior of the server on connection loss. If true, an exception is thrown immediately if the connection between client and server was lost. If false, the ClickHouse server retries to execute the query three times before throwing an exception. Note that retrying leads to increased response times. Default value: false.

• query – The custom query. Optional parameter.

:::note The table or where fields cannot be used together with the query field. And either one of the table or query fields must be declared. :::

:::note There is no explicit parameter secure. When establishing an SSL-connection security is mandatory. :::

MySQL can be connected to on a local host via sockets. To do this, set host and socket.

Example of settings:

<source>
  <mysql>
      <host>localhost</host>
      <socket>/path/to/socket/file.sock</socket>
      <user>clickhouse</user>
      <password>qwerty</password>
      <db>db_name</db>
      <table>table_name</table>
      <where>id=10</where>
      <invalidate_query>SQL_QUERY</invalidate_query>
      <fail_on_connection_loss>true</fail_on_connection_loss>
	  <query>SELECT id, value_1, value_2 FROM db_name.table_name</query>
  </mysql>
</source>

or

SOURCE(MYSQL(
    host 'localhost'
    socket '/path/to/socket/file.sock'
    user 'clickhouse'
    password 'qwerty'
    db 'db_name'
    table 'table_name'
    where 'id=10'
    invalidate_query 'SQL_QUERY'
    fail_on_connection_loss 'true'
	query 'SELECT id, value_1, value_2 FROM db_name.table_name'
))

ClickHouse

Example of settings:

<source>
    <clickhouse>
        <host>example01-01-1</host>
        <port>9000</port>
        <user>default</user>
        <password></password>
        <db>default</db>
        <table>ids</table>
        <where>id=10</where>
        <secure>1</secure>
		<query>SELECT id, value_1, value_2 FROM default.ids</query>
    </clickhouse>
</source>

or

SOURCE(CLICKHOUSE(
    host 'example01-01-1'
    port 9000
    user 'default'
    password ''
    db 'default'
    table 'ids'
    where 'id=10'
    secure 1
	query 'SELECT id, value_1, value_2 FROM default.ids'
));

Setting fields:

• host – The ClickHouse host. If it is a local host, the query is processed without any network activity. To improve fault tolerance, you can create a Distributed table and enter it in subsequent configurations.
• port – The port on the ClickHouse server.
• user – Name of the ClickHouse user.
• password – Password of the ClickHouse user.
• db – Name of the database.
• table – Name of the table.
• where – The selection criteria. May be omitted.
• invalidate_query – Query for checking the dictionary status. Optional parameter. Read more in the section Refreshing dictionary data using LIFETIME.
• secure - Use ssl for connection.
• query – The custom query. Optional parameter.

:::note The table or where fields cannot be used together with the query field. And either one of the table or query fields must be declared. :::

MongoDB

Example of settings:

<source>
    <mongodb>
        <host>localhost</host>
        <port>27017</port>
        <user></user>
        <password></password>
        <db>test</db>
        <collection>dictionary_source</collection>
        <options>ssl=true</options>
    </mongodb>
</source>

or

<source>
    <mongodb>
        <uri>mongodb://localhost:27017/test?ssl=true</uri>
        <collection>dictionary_source</collection>
    </mongodb>
</source>

or

SOURCE(MONGODB(
    host 'localhost'
    port 27017
    user ''
    password ''
    db 'test'
    collection 'dictionary_source'
    options 'ssl=true'
))

Setting fields:

• host – The MongoDB host.
• port – The port on the MongoDB server.
• user – Name of the MongoDB user.
• password – Password of the MongoDB user.
• db – Name of the database.
• collection – Name of the collection.
• options - MongoDB connection string options (optional parameter).

or

SOURCE(MONGODB(
    uri 'mongodb://localhost:27017/clickhouse'
    collection 'dictionary_source'
))

Setting fields:

• uri - URI for establish the connection.
• collection – Name of the collection.

More information about the engine

Redis

Example of settings:

<source>
    <redis>
        <host>localhost</host>
        <port>6379</port>
        <storage_type>simple</storage_type>
        <db_index>0</db_index>
    </redis>
</source>

or

SOURCE(REDIS(
    host 'localhost'
    port 6379
    storage_type 'simple'
    db_index 0
))

Setting fields:

• host – The Redis host.
• port – The port on the Redis server.
• storage_type – The structure of internal Redis storage using for work with keys. simple is for simple sources and for hashed single key sources, hash_map is for hashed sources with two keys. Ranged sources and cache sources with complex key are unsupported. May be omitted, default value is simple.
• db_index – The specific numeric index of Redis logical database. May be omitted, default value is 0.

Cassandra

Example of settings:

<source>
    <cassandra>
        <host>localhost</host>
        <port>9042</port>
        <user>username</user>
        <password>qwerty123</password>
        <keyspase>database_name</keyspase>
        <column_family>table_name</column_family>
        <allow_filtering>1</allow_filtering>
        <partition_key_prefix>1</partition_key_prefix>
        <consistency>One</consistency>
        <where>"SomeColumn" = 42</where>
        <max_threads>8</max_threads>
        <query>SELECT id, value_1, value_2 FROM database_name.table_name</query>
    </cassandra>
</source>

Setting fields:

• host – The Cassandra host or comma-separated list of hosts.
• port – The port on the Cassandra servers. If not specified, default port 9042 is used.
• user – Name of the Cassandra user.
• password – Password of the Cassandra user.
• keyspace – Name of the keyspace (database).
• column_family – Name of the column family (table).
• allow_filtering – Flag to allow or not potentially expensive conditions on clustering key columns. Default value is 1.
• partition_key_prefix – Number of partition key columns in primary key of the Cassandra table. Required for compose key dictionaries. Order of key columns in the dictionary definition must be the same as in Cassandra. Default value is 1 (the first key column is a partition key and other key columns are clustering key).
• consistency – Consistency level. Possible values: One, Two, Three, All, EachQuorum, Quorum, LocalQuorum, LocalOne, Serial, LocalSerial. Default value is One.
• where – Optional selection criteria.
• max_threads – The maximum number of threads to use for loading data from multiple partitions in compose key dictionaries.
• query – The custom query. Optional parameter.

:::note The column_family or where fields cannot be used together with the query field. And either one of the column_family or query fields must be declared. :::

PostgreSQL

Example of settings:

<source>
  <postgresql>
      <host>postgresql-hostname</hoat>
      <port>5432</port>
      <user>clickhouse</user>
      <password>qwerty</password>
      <db>db_name</db>
      <table>table_name</table>
      <where>id=10</where>
      <invalidate_query>SQL_QUERY</invalidate_query>
      <query>SELECT id, value_1, value_2 FROM db_name.table_name</query>
  </postgresql>
</source>

or

SOURCE(POSTGRESQL(
    port 5432
    host 'postgresql-hostname'
    user 'postgres_user'
    password 'postgres_password'
    db 'db_name'
    table 'table_name'
    replica(host 'example01-1' port 5432 priority 1)
    replica(host 'example01-2' port 5432 priority 2)
    where 'id=10'
    invalidate_query 'SQL_QUERY'
    query 'SELECT id, value_1, value_2 FROM db_name.table_name'
))

Setting fields:

• host – The host on the PostgreSQL server. You can specify it for all replicas, or for each one individually (inside <replica>).
• port – The port on the PostgreSQL server. You can specify it for all replicas, or for each one individually (inside <replica>).
• user – Name of the PostgreSQL user. You can specify it for all replicas, or for each one individually (inside <replica>).
• password – Password of the PostgreSQL user. You can specify it for all replicas, or for each one individually (inside <replica>).
• replica – Section of replica configurations. There can be multiple sections:
  • replica/host – The PostgreSQL host.
  • replica/port – The PostgreSQL port.
  • replica/priority – The replica priority. When attempting to connect, ClickHouse traverses the replicas in order of priority. The lower the number, the higher the priority.
• db – Name of the database.
• table – Name of the table.
• where – The selection criteria. The syntax for conditions is the same as for WHERE clause in PostgreSQL. For example, id > 10 AND id < 20. Optional parameter.
• invalidate_query – Query for checking the dictionary status. Optional parameter. Read more in the section Refreshing dictionary data using LIFETIME.
• query – The custom query. Optional parameter.

:::note The table or where fields cannot be used together with the query field. And either one of the table or query fields must be declared. :::

Null

A special source that can be used to create dummy (empty) dictionaries. Such dictionaries can useful for tests or with setups with separated data and query nodes at nodes with Distributed tables.

CREATE DICTIONARY null_dict (
    id              UInt64,
    val             UInt8,
    default_val     UInt8 DEFAULT 123,
    nullable_val    Nullable(UInt8)
)
PRIMARY KEY id
SOURCE(NULL())
LAYOUT(FLAT())
LIFETIME(0);

Dictionary Key and Fields

The structure clause describes the dictionary key and fields available for queries.

XML description:

<dictionary>
    <structure>
        <id>
            <name>Id</name>
        </id>

        <attribute>
            <!-- Attribute parameters -->
        </attribute>

        ...

    </structure>
</dictionary>

Attributes are described in the elements:

• <id> — Key column
• <attribute> — Data column: there can be a multiple number of attributes.

DDL query:

CREATE DICTIONARY dict_name (
    Id UInt64,
    -- attributes
)
PRIMARY KEY Id
...

Attributes are described in the query body:

• PRIMARY KEY — Key column
• AttrName AttrType — Data column. There can be a multiple number of attributes.

Key

ClickHouse supports the following types of keys:

• Numeric key. UInt64. Defined in the <id> tag or using PRIMARY KEY keyword.
• Composite key. Set of values of different types. Defined in the tag <key> or PRIMARY KEY keyword.

An xml structure can contain either <id> or <key>. DDL-query must contain single PRIMARY KEY.

:::note You must not describe key as an attribute. :::

Numeric Key

Type: UInt64.

Configuration example:

<id>
    <name>Id</name>
</id>

Configuration fields:

• name – The name of the column with keys.

For DDL-query:

CREATE DICTIONARY (
    Id UInt64,
    ...
)
PRIMARY KEY Id
...

• PRIMARY KEY – The name of the column with keys.

Composite Key

The key can be a tuple from any types of fields. The layout in this case must be complex_key_hashed or complex_key_cache.

:::tip A composite key can consist of a single element. This makes it possible to use a string as the key, for instance. :::

The key structure is set in the element <key>. Key fields are specified in the same format as the dictionary attributes. Example:

<structure>
    <key>
        <attribute>
            <name>field1</name>
            <type>String</type>
        </attribute>
        <attribute>
            <name>field2</name>
            <type>UInt32</type>
        </attribute>
        ...
    </key>
...

or

CREATE DICTIONARY (
    field1 String,
    field2 String
    ...
)
PRIMARY KEY field1, field2
...

For a query to the dictGet* function, a tuple is passed as the key. Example: dictGetString('dict_name', 'attr_name', tuple('string for field1', num_for_field2)).

Attributes

Configuration example:

<structure>
    ...
    <attribute>
        <name>Name</name>
        <type>ClickHouseDataType</type>
        <null_value></null_value>
        <expression>rand64()</expression>
        <hierarchical>true</hierarchical>
        <injective>true</injective>
        <is_object_id>true</is_object_id>
    </attribute>
</structure>

or

CREATE DICTIONARY somename (
    Name ClickHouseDataType DEFAULT '' EXPRESSION rand64() HIERARCHICAL INJECTIVE IS_OBJECT_ID
)

Configuration fields:

Tag	Description	Required
name	Column name.	Yes
type	ClickHouse data type: UInt8, UInt16, UInt32, UInt64, Int8, Int16, Int32, Int64, Float32, Float64, UUID, Decimal32, Decimal64, Decimal128, Decimal256,Date, Date32, DateTime, DateTime64, String, Array. ClickHouse tries to cast value from dictionary to the specified data type. For example, for MySQL, the field might be TEXT, VARCHAR, or BLOB in the MySQL source table, but it can be uploaded as String in ClickHouse. Nullable is currently supported for Flat, Hashed, ComplexKeyHashed, Direct, ComplexKeyDirect, RangeHashed, Polygon, Cache, ComplexKeyCache, SSDCache, SSDComplexKeyCache dictionaries. In IPTrie dictionaries Nullable types are not supported.	Yes
null_value	Default value for a non-existing element. In the example, it is an empty string. NULL value can be used only for the Nullable types (see the previous line with types description).	Yes
expression	Expression that ClickHouse executes on the value. The expression can be a column name in the remote SQL database. Thus, you can use it to create an alias for the remote column. Default value: no expression.	No
hierarchical	If true, the attribute contains the value of a parent key for the current key. See Hierarchical Dictionaries. Default value: false.	No
injective	Flag that shows whether the id -> attribute image is injective. If true, ClickHouse can automatically place after the GROUP BY clause the requests to dictionaries with injection. Usually it significantly reduces the amount of such requests. Default value: false.	No
is_object_id	Flag that shows whether the query is executed for a MongoDB document by ObjectID. Default value: false.

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 dictionary can be the following:

<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>

Polygon dictionaries

Polygon dictionaries allow you to efficiently search for the polygon containing specified points. For example: defining a city area by geographical coordinates.

Example of a polygon dictionary configuration:

<dictionary>
    <structure>
        <key>
            <attribute>
                <name>key</name>
                <type>Array(Array(Array(Array(Float64))))</type>
            </attribute>
        </key>

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

        <attribute>
            <name>value</name>
            <type>UInt64</type>
            <null_value>0</null_value>
        </attribute>
    </structure>

    <layout>
        <polygon>
            <store_polygon_key_column>1</store_polygon_key_column>
        </polygon>
    </layout>

    ...
</dictionary>

The corresponding DDL-query:

CREATE DICTIONARY polygon_dict_name (
    key Array(Array(Array(Array(Float64)))),
    name String,
    value UInt64
)
PRIMARY KEY key
LAYOUT(POLYGON(STORE_POLYGON_KEY_COLUMN 1))
...

When configuring the polygon dictionary, the key must have one of two types:

• A simple polygon. It is an array of points.
• MultiPolygon. It is an array of polygons. Each polygon is a two-dimensional array of points. The first element of this array is the outer boundary of the polygon, and subsequent elements specify areas to be excluded from it.

Points can be specified as an array or a tuple of their coordinates. In the current implementation, only two-dimensional points are supported.

The user can upload their own data in all formats supported by ClickHouse.

There are 3 types of in-memory storage available:

• POLYGON_SIMPLE. This is a naive implementation, where a linear pass through all polygons is made for each query, and membership is checked for each one without using additional indexes.

• POLYGON_INDEX_EACH. A separate index is built for each polygon, which allows you to quickly check whether it belongs in most cases (optimized for geographical regions). Also, a grid is superimposed on the area under consideration, which significantly narrows the number of polygons under consideration. The grid is created by recursively dividing the cell into 16 equal parts and is configured with two parameters. The division stops when the recursion depth reaches MAX_DEPTH or when the cell crosses no more than MIN_INTERSECTIONS polygons. To respond to the query, there is a corresponding cell, and the index for the polygons stored in it is accessed alternately.

• POLYGON_INDEX_CELL. This placement also creates the grid described above. The same options are available. For each sheet cell, an index is built on all pieces of polygons that fall into it, which allows you to quickly respond to a request.

• POLYGON. Synonym to POLYGON_INDEX_CELL.

Dictionary queries are carried out using standard functions for working with dictionaries. An important difference is that here the keys will be the points for which you want to find the polygon containing them.

Example

Example of working with the dictionary defined above:

CREATE TABLE points (
    x Float64,
    y Float64
)
...
SELECT tuple(x, y) AS key, dictGet(dict_name, 'name', key), dictGet(dict_name, 'value', key) FROM points ORDER BY x, y;

As a result of executing the last command for each point in the 'points' table, a minimum area polygon containing this point will be found, and the requested attributes will be output.

Example

You can read columns from polygon dictionaries via SELECT query, just turn on the store_polygon_key_column = 1 in the dictionary configuration or corresponding DDL-query.

Query:

CREATE TABLE polygons_test_table
(
    key Array(Array(Array(Tuple(Float64, Float64)))),
    name String
) ENGINE = TinyLog;

INSERT INTO polygons_test_table VALUES ([[[(3, 1), (0, 1), (0, -1), (3, -1)]]], 'Value');

CREATE DICTIONARY polygons_test_dictionary
(
    key Array(Array(Array(Tuple(Float64, Float64)))),
    name String
)
PRIMARY KEY key
SOURCE(CLICKHOUSE(TABLE 'polygons_test_table'))
LAYOUT(POLYGON(STORE_POLYGON_KEY_COLUMN 1))
LIFETIME(0);

SELECT * FROM polygons_test_dictionary;

Result:

┌─key─────────────────────────────┬─name──┐
│ [[[(3,1),(0,1),(0,-1),(3,-1)]]] │ Value │
└─────────────────────────────────┴───────┘

Regular Expression Tree Dictionary

Regular expression tree dictionaries are a special type of dictionary which represent the mapping from key to attributes using a tree of regular expressions. There are some use cases, e.g. parsing of user agent strings, which can be expressed elegantly with regexp tree dictionaries.

Use Regular Expression Tree Dictionary in ClickHouse Open-Source

Regular expression tree dictionaries are defined in ClickHouse open-source using the YAMLRegExpTree source which is provided the path to a YAML file containing the regular expression tree.

CREATE DICTIONARY regexp_dict
(
    regexp String,
    name String,
    version String
)
PRIMARY KEY(regexp)
SOURCE(YAMLRegExpTree(PATH '/var/lib/clickhouse/user_files/regexp_tree.yaml'))
LAYOUT(regexp_tree)
...

The dictionary source YAMLRegExpTree represents the structure of a regexp tree. For example:

- regexp: 'Linux/(\d+[\.\d]*).+tlinux'
  name: 'TencentOS'
  version: '\1'

- regexp: '\d+/tclwebkit(?:\d+[\.\d]*)'
  name: 'Android'
  versions:
    - regexp: '33/tclwebkit'
      version: '13'
    - regexp: '3[12]/tclwebkit'
      version: '12'
    - regexp: '30/tclwebkit'
      version: '11'
    - regexp: '29/tclwebkit'
      version: '10'

This config consists of a list of regular expression tree nodes. Each node has the following structure:

• regexp: the regular expression of the node.
• attributes: a list of user-defined dictionary attributes. In this example, there are two attributes: name and version. The first node defines both attributes. The second node only defines attribute name. Attribute version is provided by the child nodes of the second node.
  • The value of an attribute may contain back references, referring to capture groups of the matched regular expression. In the example, the value of attribute version in the first node consists of a back-reference \1 to capture group (\d+[\.\d]*) in the regular expression. Back-reference numbers range from 1 to 9 and are written as $1 or \1 (for number 1). The back reference is replaced by the matched capture group during query execution.
• child nodes: a list of children of a regexp tree node, each of which has its own attributes and (potentially) children nodes. String matching proceeds in a depth-first fashion. If a string matches a regexp node, the dictionary checks if it also matches the nodes' child nodes. If that is the case, the attributes of the deepest matching node are assigned. Attributes of a child node overwrite equally named attributes of parent nodes. The name of child nodes in YAML files can be arbitrary, e.g. versions in above example.

Regexp tree dictionaries only allow access using the functions dictGet, dictGetOrDefault, and dictGetAll.

Example:

SELECT dictGet('regexp_dict', ('name', 'version'), '31/tclwebkit1024');

Result:

┌─dictGet('regexp_dict', ('name', 'version'), '31/tclwebkit1024')─┐
│ ('Android','12')                                                │
└─────────────────────────────────────────────────────────────────┘

In this case, we first match the regular expression \d+/tclwebkit(?:\d+[\.\d]*) in the top layer's second node. The dictionary then continues to look into the child nodes and finds that the string also matches 3[12]/tclwebkit. As a result, the value of attribute name is Android (defined in the first layer) and the value of attribute version is 12 (defined the child node).

With a powerful YAML configure file, we can use a regexp tree dictionaries as a user agent string parser. We support uap-core and demonstrate how to use it in the functional test 02504_regexp_dictionary_ua_parser

Collecting Attribute Values

Sometimes it is useful to return values from multiple regular expressions that matched, rather than just the value of a leaf node. In these cases, the specialized dictGetAll function can be used. If a node has an attribute value of type T, dictGetAll will return an Array(T) containing zero or more values.

By default, the number of matches returned per key is unbounded. A bound can be passed as an optional fourth argument to dictGetAll. The array is populated in topological order, meaning that child nodes come before parent nodes, and sibling nodes follow the ordering in the source.

Example:

CREATE DICTIONARY regexp_dict
(
    regexp String,
    tag String,
    topological_index Int64,
    captured Nullable(String),
    parent String
)
PRIMARY KEY(regexp)
SOURCE(YAMLRegExpTree(PATH '/var/lib/clickhouse/user_files/regexp_tree.yaml'))
LAYOUT(regexp_tree)
LIFETIME(0)

# /var/lib/clickhouse/user_files/regexp_tree.yaml
- regexp: 'clickhouse\.com'
  tag: 'ClickHouse'
  topological_index: 1
  paths:
    - regexp: 'clickhouse\.com/docs(.*)'
      tag: 'ClickHouse Documentation'
      topological_index: 0
      captured: '\1'
      parent: 'ClickHouse'

- regexp: '/docs(/|$)'
  tag: 'Documentation'
  topological_index: 2

- regexp: 'github.com'
  tag: 'GitHub'
  topological_index: 3
  captured: 'NULL'

CREATE TABLE urls (url String) ENGINE=MergeTree ORDER BY url;
INSERT INTO urls VALUES ('clickhouse.com'), ('clickhouse.com/docs/en'), ('github.com/clickhouse/tree/master/docs');
SELECT url, dictGetAll('regexp_dict', ('tag', 'topological_index', 'captured', 'parent'), url, 2) FROM urls;

Result:

┌─url────────────────────────────────────┬─dictGetAll('regexp_dict', ('tag', 'topological_index', 'captured', 'parent'), url, 2)─┐
│ clickhouse.com                         │ (['ClickHouse'],[1],[],[])                                                            │
│ clickhouse.com/docs/en                 │ (['ClickHouse Documentation','ClickHouse'],[0,1],['/en'],['ClickHouse'])              │
│ github.com/clickhouse/tree/master/docs │ (['Documentation','GitHub'],[2,3],[NULL],[])                                          │
└────────────────────────────────────────┴───────────────────────────────────────────────────────────────────────────────────────┘

Matching Modes

Pattern matching behavior can be modified with certain dictionary settings:

• regexp_dict_flag_case_insensitive: Use case-insensitive matching (defaults to false). Can be overridden in individual expressions with (?i) and (?-i).
• regexp_dict_flag_dotall: Allow '.' to match newline characters (defaults to false).

Use Regular Expression Tree Dictionary in ClickHouse Cloud

Above used YAMLRegExpTree source works in ClickHouse Open Source but not in ClickHouse Cloud. To use regexp tree dictionaries in ClickHouse could, first create a regexp tree dictionary from a YAML file locally in ClickHouse Open Source, then dump this dictionary into a CSV file using the dictionary table function and the INTO OUTFILE clause.

SELECT * FROM dictionary(regexp_dict) INTO OUTFILE('regexp_dict.csv')

The content of csv file is:

1,0,"Linux/(\d+[\.\d]*).+tlinux","['version','name']","['\\1','TencentOS']"
2,0,"(\d+)/tclwebkit(\d+[\.\d]*)","['comment','version','name']","['test $1 and $2','$1','Android']"
3,2,"33/tclwebkit","['version']","['13']"
4,2,"3[12]/tclwebkit","['version']","['12']"
5,2,"3[12]/tclwebkit","['version']","['11']"
6,2,"3[12]/tclwebkit","['version']","['10']"

The schema of dumped file is:

• id UInt64: the id of the RegexpTree node.
• parent_id UInt64: the id of the parent of a node.
• regexp String: the regular expression string.
• keys Array(String): the names of user-defined attributes.
• values Array(String): the values of user-defined attributes.

To create the dictionary in ClickHouse Cloud, first create a table regexp_dictionary_source_table with below table structure:

CREATE TABLE regexp_dictionary_source_table
(
    id UInt64,
    parent_id UInt64,
    regexp String,
    keys   Array(String),
    values Array(String)
) ENGINE=Memory;

Then update the local CSV by

clickhouse client \
    --host MY_HOST \
    --secure \
    --password MY_PASSWORD \
    --query "
    INSERT INTO regexp_dictionary_source_table
    SELECT * FROM input ('id UInt64, parent_id UInt64, regexp String, keys Array(String), values Array(String)')
    FORMAT CSV" < regexp_dict.csv

You can see how to Insert Local Files for more details. After we initialize the source table, we can create a RegexpTree by table source:

CREATE DICTIONARY regexp_dict
(
    regexp String,
    name String,
    version String
PRIMARY KEY(regexp)
SOURCE(CLICKHOUSE(TABLE 'regexp_dictionary_source_table'))
LIFETIME(0)
LAYOUT(regexp_tree);

Embedded Dictionaries

ClickHouse contains a built-in feature for working with a geobase.

This allows you to:

• Use a region’s ID to get its name in the desired language.
• Use a region’s ID to get the ID of a city, area, federal district, country, or continent.
• Check whether a region is part of another region.
• Get a chain of parent regions.

All the functions support “translocality,” the ability to simultaneously use different perspectives on region ownership. For more information, see the section “Functions for working with web analytics dictionaries”.

The internal dictionaries are disabled in the default package. To enable them, uncomment the parameters path_to_regions_hierarchy_file and path_to_regions_names_files in the server configuration file.

The geobase is loaded from text files.

Place the regions_hierarchy*.txt files into the path_to_regions_hierarchy_file directory. This configuration parameter must contain the path to the regions_hierarchy.txt file (the default regional hierarchy), and the other files (regions_hierarchy_ua.txt) must be located in the same directory.

Put the regions_names_*.txt files in the path_to_regions_names_files directory.

You can also create these files yourself. The file format is as follows:

regions_hierarchy*.txt: TabSeparated (no header), columns:

• region ID (UInt32)
• parent region ID (UInt32)
• region type (UInt8): 1 - continent, 3 - country, 4 - federal district, 5 - region, 6 - city; other types do not have values
• population (UInt32) — optional column

regions_names_*.txt: TabSeparated (no header), columns:

• region ID (UInt32)
• region name (String) — Can’t contain tabs or line feeds, even escaped ones.

A flat array is used for storing in RAM. For this reason, IDs shouldn’t be more than a million.

Dictionaries can be updated without restarting the server. However, the set of available dictionaries is not updated. For updates, the file modification times are checked. If a file has changed, the dictionary is updated. The interval to check for changes is configured in the builtin_dictionaries_reload_interval parameter. Dictionary updates (other than loading at first use) do not block queries. During updates, queries use the old versions of dictionaries. If an error occurs during an update, the error is written to the server log, and queries continue using the old version of dictionaries.

We recommend periodically updating the dictionaries with the geobase. During an update, generate new files and write them to a separate location. When everything is ready, rename them to the files used by the server.

There are also functions for working with OS identifiers and search engines, but they shouldn’t be used.