2020-04-03 13:23:32 +00:00
---
toc_priority: 56
2020-06-19 10:16:30 +00:00
toc_title: JSON
2020-04-03 13:23:32 +00:00
---
2020-04-30 18:19:18 +00:00
# Functions for Working with JSON {#functions-for-working-with-json}
2017-12-28 15:13:23 +00:00
In Yandex.Metrica, JSON is transmitted by users as session parameters. There are some special functions for working with this JSON. (Although in most of the cases, the JSONs are additionally pre-processed, and the resulting values are put in separate columns in their processed format.) All these functions are based on strong assumptions about what the JSON can be, but they try to do as little as possible to get the job done.
The following assumptions are made:
2020-03-20 10:10:48 +00:00
1. The field name (function argument) must be a constant.
2. The field name is somehow canonically encoded in JSON. For example: `visitParamHas('{"abc":"def"}', 'abc') = 1` , but `visitParamHas('{"\\u0061\\u0062\\u0063":"def"}', 'abc') = 0`
3. Fields are searched for on any nesting level, indiscriminately. If there are multiple matching fields, the first occurrence is used.
2021-05-27 19:44:11 +00:00
4. The JSON does not have space characters outside of string literals.
2017-12-28 15:13:23 +00:00
2020-03-20 10:10:48 +00:00
## visitParamHas(params, name) {#visitparamhasparams-name}
2017-12-28 15:13:23 +00:00
2021-04-17 17:04:32 +00:00
Checks whether there is a field with the `name` name.
Alias: `simpleJSONHas` .
2017-12-28 15:13:23 +00:00
2020-03-20 10:10:48 +00:00
## visitParamExtractUInt(params, name) {#visitparamextractuintparams-name}
2017-12-28 15:13:23 +00:00
2021-05-27 19:44:11 +00:00
Parses UInt64 from the value of the field named `name` . If this is a string field, it tries to parse a number from the beginning of the string. If the field does not exist, or it exists but does not contain a number, it returns 0.
2021-04-17 17:04:32 +00:00
Alias: `simpleJSONExtractUInt` .
2017-12-28 15:13:23 +00:00
2020-03-20 10:10:48 +00:00
## visitParamExtractInt(params, name) {#visitparamextractintparams-name}
2017-12-28 15:13:23 +00:00
The same as for Int64.
2021-04-17 17:04:32 +00:00
Alias: `simpleJSONExtractInt` .
2020-03-20 10:10:48 +00:00
## visitParamExtractFloat(params, name) {#visitparamextractfloatparams-name}
2017-12-28 15:13:23 +00:00
The same as for Float64.
2021-04-17 17:04:32 +00:00
Alias: `simpleJSONExtractFloat` .
2020-03-20 10:10:48 +00:00
## visitParamExtractBool(params, name) {#visitparamextractboolparams-name}
2017-12-28 15:13:23 +00:00
Parses a true/false value. The result is UInt8.
2021-04-17 17:04:32 +00:00
Alias: `simpleJSONExtractBool` .
2020-03-20 10:10:48 +00:00
## visitParamExtractRaw(params, name) {#visitparamextractrawparams-name}
2017-12-28 15:13:23 +00:00
Returns the value of a field, including separators.
2021-04-17 17:04:32 +00:00
Alias: `simpleJSONExtractRaw` .
2017-12-28 15:13:23 +00:00
Examples:
2020-03-20 10:10:48 +00:00
``` sql
2021-04-17 17:04:32 +00:00
visitParamExtractRaw('{"abc":"\\n\\u0000"}', 'abc') = '"\\n\\u0000"';
visitParamExtractRaw('{"abc":{"def":[1,2,3]}}', 'abc') = '{"def":[1,2,3]}';
2017-12-28 15:13:23 +00:00
```
2020-03-20 10:10:48 +00:00
## visitParamExtractString(params, name) {#visitparamextractstringparams-name}
2017-12-28 15:13:23 +00:00
Parses the string in double quotes. The value is unescaped. If unescaping failed, it returns an empty string.
2021-04-17 17:04:32 +00:00
Alias: `simpleJSONExtractString` .
2017-12-28 15:13:23 +00:00
Examples:
2020-03-20 10:10:48 +00:00
``` sql
2021-04-17 17:04:32 +00:00
visitParamExtractString('{"abc":"\\n\\u0000"}', 'abc') = '\n\0';
visitParamExtractString('{"abc":"\\u263a"}', 'abc') = '☺';
visitParamExtractString('{"abc":"\\u263"}', 'abc') = '';
visitParamExtractString('{"abc":"hello}', 'abc') = '';
2017-12-28 15:13:23 +00:00
```
There is currently no support for code points in the format `\uXXXX\uYYYY` that are not from the basic multilingual plane (they are converted to CESU-8 instead of UTF-8).
2019-04-24 04:23:14 +00:00
The following functions are based on [simdjson ](https://github.com/lemire/simdjson ) designed for more complex JSON parsing requirements. The assumption 2 mentioned above still applies.
2020-03-20 10:10:48 +00:00
## isValidJSON(json) {#isvalidjsonjson}
2019-10-12 12:00:46 +00:00
Checks that passed string is a valid json.
Examples:
2020-03-20 10:10:48 +00:00
``` sql
2019-10-12 12:00:46 +00:00
SELECT isValidJSON('{"a": "hello", "b": [-100, 200.0, 300]}') = 1
SELECT isValidJSON('not a json') = 0
```
2020-10-13 17:23:29 +00:00
## JSONHas(json\[, indices_or_keys\]…) {#jsonhasjson-indices-or-keys}
2019-04-24 04:23:14 +00:00
If the value exists in the JSON document, `1` will be returned.
2019-05-16 13:35:31 +00:00
If the value does not exist, `0` will be returned.
2019-04-24 04:23:14 +00:00
Examples:
2020-03-20 10:10:48 +00:00
``` sql
2019-09-23 15:31:46 +00:00
SELECT JSONHas('{"a": "hello", "b": [-100, 200.0, 300]}', 'b') = 1
SELECT JSONHas('{"a": "hello", "b": [-100, 200.0, 300]}', 'b', 4) = 0
2019-04-24 04:23:14 +00:00
```
2019-05-16 13:35:31 +00:00
`indices_or_keys` is a list of zero or more arguments each of them can be either string or integer.
2019-04-24 04:23:14 +00:00
2020-03-21 04:11:51 +00:00
- String = access object member by key.
- Positive integer = access the n-th member/key from the beginning.
- Negative integer = access the n-th member/key from the end.
2019-04-24 04:23:14 +00:00
2021-05-27 19:44:11 +00:00
Minimum index of the element is 1. Thus the element 0 does not exist.
2019-06-14 12:44:33 +00:00
2019-05-10 08:49:03 +00:00
You may use integers to access both JSON arrays and JSON objects.
2019-04-24 04:23:14 +00:00
So, for example:
2020-03-20 10:10:48 +00:00
``` sql
2019-09-23 15:31:46 +00:00
SELECT JSONExtractKey('{"a": "hello", "b": [-100, 200.0, 300]}', 1) = 'a'
SELECT JSONExtractKey('{"a": "hello", "b": [-100, 200.0, 300]}', 2) = 'b'
SELECT JSONExtractKey('{"a": "hello", "b": [-100, 200.0, 300]}', -1) = 'b'
SELECT JSONExtractKey('{"a": "hello", "b": [-100, 200.0, 300]}', -2) = 'a'
SELECT JSONExtractString('{"a": "hello", "b": [-100, 200.0, 300]}', 1) = 'hello'
2019-04-24 04:23:14 +00:00
```
2020-10-13 17:23:29 +00:00
## JSONLength(json\[, indices_or_keys\]…) {#jsonlengthjson-indices-or-keys}
2019-04-24 04:23:14 +00:00
2019-05-10 08:49:03 +00:00
Return the length of a JSON array or a JSON object.
2019-04-24 04:23:14 +00:00
2019-05-16 13:35:31 +00:00
If the value does not exist or has a wrong type, `0` will be returned.
2019-04-24 04:23:14 +00:00
Examples:
2020-03-20 10:10:48 +00:00
``` sql
2019-09-23 15:31:46 +00:00
SELECT JSONLength('{"a": "hello", "b": [-100, 200.0, 300]}', 'b') = 3
SELECT JSONLength('{"a": "hello", "b": [-100, 200.0, 300]}') = 2
2019-04-24 04:23:14 +00:00
```
2020-10-13 17:23:29 +00:00
## JSONType(json\[, indices_or_keys\]…) {#jsontypejson-indices-or-keys}
2019-04-24 04:23:14 +00:00
Return the type of a JSON value.
2019-05-16 13:35:31 +00:00
If the value does not exist, `Null` will be returned.
2019-04-24 04:23:14 +00:00
Examples:
2020-03-20 10:10:48 +00:00
``` sql
2019-09-23 15:31:46 +00:00
SELECT JSONType('{"a": "hello", "b": [-100, 200.0, 300]}') = 'Object'
SELECT JSONType('{"a": "hello", "b": [-100, 200.0, 300]}', 'a') = 'String'
SELECT JSONType('{"a": "hello", "b": [-100, 200.0, 300]}', 'b') = 'Array'
2019-04-24 04:23:14 +00:00
```
2020-10-13 17:23:29 +00:00
## JSONExtractUInt(json\[, indices_or_keys\]…) {#jsonextractuintjson-indices-or-keys}
2020-03-20 10:10:48 +00:00
2020-10-13 17:23:29 +00:00
## JSONExtractInt(json\[, indices_or_keys\]…) {#jsonextractintjson-indices-or-keys}
2020-03-20 10:10:48 +00:00
2020-10-13 17:23:29 +00:00
## JSONExtractFloat(json\[, indices_or_keys\]…) {#jsonextractfloatjson-indices-or-keys}
2020-03-20 10:10:48 +00:00
2020-10-13 17:23:29 +00:00
## JSONExtractBool(json\[, indices_or_keys\]…) {#jsonextractbooljson-indices-or-keys}
2019-04-24 04:23:14 +00:00
2019-05-16 13:35:31 +00:00
Parses a JSON and extract a value. These functions are similar to `visitParam` functions.
2019-04-24 04:23:14 +00:00
2019-05-16 13:35:31 +00:00
If the value does not exist or has a wrong type, `0` will be returned.
2019-04-24 04:23:14 +00:00
Examples:
2020-03-20 10:10:48 +00:00
``` sql
2019-09-23 15:31:46 +00:00
SELECT JSONExtractInt('{"a": "hello", "b": [-100, 200.0, 300]}', 'b', 1) = -100
SELECT JSONExtractFloat('{"a": "hello", "b": [-100, 200.0, 300]}', 'b', 2) = 200.0
SELECT JSONExtractUInt('{"a": "hello", "b": [-100, 200.0, 300]}', 'b', -1) = 300
2019-04-24 04:23:14 +00:00
```
2020-10-13 17:23:29 +00:00
## JSONExtractString(json\[, indices_or_keys\]…) {#jsonextractstringjson-indices-or-keys}
2019-04-24 04:23:14 +00:00
2019-05-16 13:35:31 +00:00
Parses a JSON and extract a string. This function is similar to `visitParamExtractString` functions.
2019-04-24 04:23:14 +00:00
2019-05-16 13:35:31 +00:00
If the value does not exist or has a wrong type, an empty string will be returned.
2019-04-24 04:23:14 +00:00
2019-05-16 13:35:31 +00:00
The value is unescaped. If unescaping failed, it returns an empty string.
2019-04-24 04:23:14 +00:00
Examples:
2020-03-20 10:10:48 +00:00
``` sql
2019-09-23 15:31:46 +00:00
SELECT JSONExtractString('{"a": "hello", "b": [-100, 200.0, 300]}', 'a') = 'hello'
SELECT JSONExtractString('{"abc":"\\n\\u0000"}', 'abc') = '\n\0'
SELECT JSONExtractString('{"abc":"\\u263a"}', 'abc') = '☺'
SELECT JSONExtractString('{"abc":"\\u263"}', 'abc') = ''
SELECT JSONExtractString('{"abc":"hello}', 'abc') = ''
2019-04-24 04:23:14 +00:00
```
2020-10-13 17:23:29 +00:00
## JSONExtract(json\[, indices_or_keys…\], Return_type) {#jsonextractjson-indices-or-keys-return-type}
2019-05-13 13:46:19 +00:00
2019-05-16 13:35:31 +00:00
Parses a JSON and extract a value of the given ClickHouse data type.
2019-05-13 13:46:19 +00:00
2019-05-16 13:35:31 +00:00
This is a generalization of the previous `JSONExtract<type>` functions.
This means
`JSONExtract(..., 'String')` returns exactly the same as `JSONExtractString()` ,
`JSONExtract(..., 'Float64')` returns exactly the same as `JSONExtractFloat()` .
2019-05-13 13:46:19 +00:00
Examples:
2020-03-20 10:10:48 +00:00
``` sql
2019-05-16 13:35:31 +00:00
SELECT JSONExtract('{"a": "hello", "b": [-100, 200.0, 300]}', 'Tuple(String, Array(Float64))') = ('hello',[-100,200,300])
SELECT JSONExtract('{"a": "hello", "b": [-100, 200.0, 300]}', 'Tuple(b Array(Float64), a String)') = ([-100,200,300],'hello')
SELECT JSONExtract('{"a": "hello", "b": [-100, 200.0, 300]}', 'b', 'Array(Nullable(Int8))') = [-100, NULL, NULL]
SELECT JSONExtract('{"a": "hello", "b": [-100, 200.0, 300]}', 'b', 4, 'Nullable(Int64)') = NULL
SELECT JSONExtract('{"passed": true}', 'passed', 'UInt8') = 1
SELECT JSONExtract('{"day": "Thursday"}', 'day', 'Enum8(\'Sunday\' = 0, \'Monday\' = 1, \'Tuesday\' = 2, \'Wednesday\' = 3, \'Thursday\' = 4, \'Friday\' = 5, \'Saturday\' = 6)') = 'Thursday'
SELECT JSONExtract('{"day": 5}', 'day', 'Enum8(\'Sunday\' = 0, \'Monday\' = 1, \'Tuesday\' = 2, \'Wednesday\' = 3, \'Thursday\' = 4, \'Friday\' = 5, \'Saturday\' = 6)') = 'Friday'
```
2020-10-13 17:23:29 +00:00
## JSONExtractKeysAndValues(json\[, indices_or_keys…\], Value_type) {#jsonextractkeysandvaluesjson-indices-or-keys-value-type}
2019-05-16 13:35:31 +00:00
2020-04-20 10:08:39 +00:00
Parses key-value pairs from a JSON where the values are of the given ClickHouse data type.
2019-05-16 13:35:31 +00:00
Example:
2020-03-20 10:10:48 +00:00
``` sql
2021-03-13 18:18:45 +00:00
SELECT JSONExtractKeysAndValues('{"x": {"a": 5, "b": 7, "c": 11}}', 'x', 'Int8') = [('a',5),('b',7),('c',11)];
2019-05-13 13:46:19 +00:00
```
2021-11-02 17:01:49 +00:00
## JSONExtractKeys(json[, indices_or_keys]…) {#jsonextractkeysjson-indices-or-keys}
2021-11-06 20:58:06 +00:00
Parses a JSON string and extracts the keys.
2021-11-02 17:01:49 +00:00
**Syntax**
``` sql
2021-11-06 20:58:15 +00:00
JSONExtractKeys(json[, a, b, c...])
2021-11-02 17:01:49 +00:00
```
**Arguments**
- `json` — [String ](../../sql-reference/data-types/string.md ) with valid JSON.
- `p, a, t, h` — Comma-separated indices or keys that specify the path to the inner field in a nested JSON object. Each argument can be either a [string ](../../sql-reference/data-types/string.md ) to get the field by the key or an [integer ](../../sql-reference/data-types/int-uint.md ) to get the N-th field (indexed from 1, negative integers count from the end). If not set, the whole JSON is parsed as the top-level object. Optional parameter.
**Returned value**
2021-11-06 20:58:24 +00:00
Array with the keys of the JSON.
2021-11-02 17:01:49 +00:00
2021-11-06 20:58:36 +00:00
Type: [Array ](../../sql-reference/data-types/array.md )([String](../../sql-reference/data-types/string.md)).
2021-11-02 17:01:49 +00:00
**Example**
Query:
```sql
SELECT JSONExtractKeys('{"a": "hello", "b": [-100, 200.0, 300]}');
```
Result:
```
text
┌─JSONExtractKeys('{"a": "hello", "b": [-100, 200.0, 300]}')─┐
│ ['a','b'] │
└────────────────────────────────────────────────────────────┘
```
2020-10-13 17:23:29 +00:00
## JSONExtractRaw(json\[, indices_or_keys\]…) {#jsonextractrawjson-indices-or-keys}
2019-05-16 13:35:31 +00:00
2020-04-20 10:08:39 +00:00
Returns a part of JSON as unparsed string.
2019-05-16 13:35:31 +00:00
If the part does not exist or has a wrong type, an empty string will be returned.
Example:
2020-03-20 10:10:48 +00:00
``` sql
2021-03-13 18:18:45 +00:00
SELECT JSONExtractRaw('{"a": "hello", "b": [-100, 200.0, 300]}', 'b') = '[-100, 200.0, 300]';
2019-05-16 13:35:31 +00:00
```
2019-06-05 10:04:00 +00:00
2020-10-13 17:23:29 +00:00
## JSONExtractArrayRaw(json\[, indices_or_keys…\]) {#jsonextractarrayrawjson-indices-or-keys}
2019-12-08 14:49:26 +00:00
2019-12-08 20:50:00 +00:00
Returns an array with elements of JSON array, each represented as unparsed string.
2019-12-08 14:49:26 +00:00
2020-03-20 10:10:48 +00:00
If the part does not exist or isn’ t array, an empty array will be returned.
2019-12-08 14:49:26 +00:00
Example:
2020-03-20 10:10:48 +00:00
``` sql
2021-03-13 18:18:45 +00:00
SELECT JSONExtractArrayRaw('{"a": "hello", "b": [-100, 200.0, "hello"]}', 'b') = ['-100', '200.0', '"hello"']';
2019-12-08 14:49:26 +00:00
```
2020-05-07 18:38:27 +00:00
## JSONExtractKeysAndValuesRaw {#json-extract-keys-and-values-raw}
2020-04-20 10:08:39 +00:00
2020-05-07 18:38:27 +00:00
Extracts raw data from a JSON object.
2020-04-20 10:08:39 +00:00
2020-05-07 18:38:27 +00:00
**Syntax**
2020-04-20 10:08:39 +00:00
2020-05-07 18:38:27 +00:00
``` sql
JSONExtractKeysAndValuesRaw(json[, p, a, t, h])
```
2021-02-15 21:22:10 +00:00
**Arguments**
2020-05-07 18:38:27 +00:00
2020-06-18 08:24:31 +00:00
- `json` — [String ](../../sql-reference/data-types/string.md ) with valid JSON.
- `p, a, t, h` — Comma-separated indices or keys that specify the path to the inner field in a nested JSON object. Each argument can be either a [string ](../../sql-reference/data-types/string.md ) to get the field by the key or an [integer ](../../sql-reference/data-types/int-uint.md ) to get the N-th field (indexed from 1, negative integers count from the end). If not set, the whole JSON is parsed as the top-level object. Optional parameter.
2020-05-07 18:38:27 +00:00
**Returned values**
2020-06-18 08:24:31 +00:00
- Array with `('key', 'value')` tuples. Both tuple members are strings.
- Empty array if the requested object does not exist, or input JSON is invalid.
2020-05-07 18:38:27 +00:00
2020-06-18 08:24:31 +00:00
Type: [Array ](../../sql-reference/data-types/array.md )([Tuple](../../sql-reference/data-types/tuple.md)([String](../../sql-reference/data-types/string.md), [String ](../../sql-reference/data-types/string.md )).
2020-05-07 18:38:27 +00:00
**Examples**
Query:
2020-04-20 10:08:39 +00:00
``` sql
2021-03-13 18:18:45 +00:00
SELECT JSONExtractKeysAndValuesRaw('{"a": [-100, 200.0], "b":{"c": {"d": "hello", "f": "world"}}}');
2020-05-07 18:38:27 +00:00
```
Result:
``` text
┌─JSONExtractKeysAndValuesRaw('{"a": [-100, 200.0], "b":{"c": {"d": "hello", "f": "world"}}}')─┐
│ [('a','[-100,200]'),('b','{"c":{"d":"hello","f":"world"}}')] │
└──────────────────────────────────────────────────────────────────────────────────────────────┘
2020-04-20 10:08:39 +00:00
```
2020-05-07 18:38:27 +00:00
Query:
``` sql
2021-03-13 18:18:45 +00:00
SELECT JSONExtractKeysAndValuesRaw('{"a": [-100, 200.0], "b":{"c": {"d": "hello", "f": "world"}}}', 'b');
2020-05-07 18:38:27 +00:00
```
Result:
``` text
┌─JSONExtractKeysAndValuesRaw('{"a": [-100, 200.0], "b":{"c": {"d": "hello", "f": "world"}}}', 'b')─┐
│ [('c','{"d":"hello","f":"world"}')] │
└───────────────────────────────────────────────────────────────────────────────────────────────────┘
```
Query:
``` sql
2021-03-13 18:18:45 +00:00
SELECT JSONExtractKeysAndValuesRaw('{"a": [-100, 200.0], "b":{"c": {"d": "hello", "f": "world"}}}', -1, 'c');
2020-05-07 18:38:27 +00:00
```
Result:
``` text
┌─JSONExtractKeysAndValuesRaw('{"a": [-100, 200.0], "b":{"c": {"d": "hello", "f": "world"}}}', -1, 'c')─┐
│ [('d','"hello"'),('f','"world"')] │
└───────────────────────────────────────────────────────────────────────────────────────────────────────┘
```
2021-07-01 20:57:01 +00:00
## toJSONString {#tojsonstring}
2021-07-06 21:01:45 +00:00
Serializes a value to its JSON representation. Various data types and nested structures are supported.
2021-07-08 20:18:56 +00:00
64-bit [integers ](../../sql-reference/data-types/int-uint.md ) or bigger (like `UInt64` or `Int128` ) are enclosed in quotes by default. [output_format_json_quote_64bit_integers ](../../operations/settings/settings.md#session_settings-output_format_json_quote_64bit_integers ) controls this behavior.
2021-07-10 07:01:34 +00:00
Special values `NaN` and `inf` are replaced with `null` . Enable [output_format_json_quote_denormals ](../../operations/settings/settings.md#settings-output_format_json_quote_denormals ) setting to show them.
2021-07-08 20:18:56 +00:00
When serializing an [Enum ](../../sql-reference/data-types/enum.md ) value, the function outputs its name.
2021-07-01 20:57:01 +00:00
**Syntax**
``` sql
toJSONString(value)
```
**Arguments**
2021-07-07 19:30:02 +00:00
- `value` — Value to serialize. Value may be of any data type.
2021-07-01 20:57:01 +00:00
**Returned value**
2021-07-29 15:20:55 +00:00
- JSON representation of the value.
2021-07-01 20:57:01 +00:00
2021-07-06 20:49:14 +00:00
Type: [String ](../../sql-reference/data-types/string.md ).
2021-07-01 20:57:01 +00:00
**Example**
2021-07-10 07:01:34 +00:00
The first example shows serialization of a [Map ](../../sql-reference/data-types/map.md ).
The second example shows some special values wrapped into a [Tuple ](../../sql-reference/data-types/tuple.md ).
2021-07-06 20:49:14 +00:00
2021-07-01 20:57:01 +00:00
Query:
``` sql
2021-07-06 20:49:14 +00:00
SELECT toJSONString(map('key1', 1, 'key2', 2));
2021-07-10 07:01:34 +00:00
SELECT toJSONString(tuple(1.25, NULL, NaN, +inf, -inf, [])) SETTINGS output_format_json_quote_denormals = 1;
2021-07-01 20:57:01 +00:00
```
Result:
``` text
2021-07-06 20:49:14 +00:00
{"key1":1,"key2":2}
2021-07-10 07:01:34 +00:00
[1.25,null,"nan","inf","-inf",[]]
2021-07-01 20:57:01 +00:00
```
2021-07-08 20:18:56 +00:00
**See Also**
- [output_format_json_quote_64bit_integers ](../../operations/settings/settings.md#session_settings-output_format_json_quote_64bit_integers )
- [output_format_json_quote_denormals ](../../operations/settings/settings.md#settings-output_format_json_quote_denormals )