JSON Functions And Operators (2024)

By default, SQLite supports twenty-nine functions and two operators fordealing with JSON values. There are also two table-valued functionsthat can be used to decompose a JSON string.

There are 25 scalar functions and operators:

1. json(json)
2. jsonb(json)
3. json_array(value1,value2,...)
4. jsonb_array(value1,value2,...)
5. json_array_length(json)
   json_array_length(json,path)
6. json_error_position(json)
7. json_extract(json,path,...)
8. jsonb_extract(json,path,...)
9. json -> path
10. json ->> path
11. json_insert(json,path,value,...)
12. jsonb_insert(json,path,value,...)
13. json_object(label1,value1,...)
14. jsonb_object(label1,value1,...)
15. json_patch(json1,json2)
16. jsonb_patch(json1,json2)
17. json_remove(json,path,...)
18. jsonb_remove(json,path,...)
19. json_replace(json,path,value,...)
20. jsonb_replace(json,path,value,...)
21. json_set(json,path,value,...)
22. jsonb_set(json,path,value,...)
23. json_type(json)
    json_type(json,path)
24. json_valid(json)
    json_valid(json,flags)
25. json_quote(value)

There are four aggregate SQL functions:

26. json_group_array(value)
27. jsonb_group_array(value)
28. json_group_object(label,value)
29. jsonb_group_object(name,value)

The two table-valued functions are:

30. json_each(json)
    json_each(json,path)
31. json_tree(json)
    json_tree(json,path)

The JSON functions and operators are built into SQLite by default,as of SQLite version 3.38.0 (2022-02-22). They can be omittedby adding the -DSQLITE_OMIT_JSON compile-time option. Prior toversion 3.38.0, the JSON functions were an extension that would onlybe included in builds if the -DSQLITE_ENABLE_JSON1 compile-time optionwas included. In other words, the JSON functions went from beingopt-in with SQLite version 3.37.2 and earlier to opt-out withSQLite version 3.38.0 and later.

SQLite stores JSON as ordinary text.Backwards compatibility constraints mean that SQLite is only able tostore values that are NULL, integers, floating-point numbers, text,and BLOBs. It is not possible to add a new "JSON" type.

3.1. JSON arguments

For functions that accept JSON as their first argument, that argumentcan be a JSON object, array, number, string, or null. SQLite numericvalues and NULL values are interpreted as JSON numbers and nulls, respectively.SQLite text values can be understood as JSON objects, arrays, or strings.If an SQLite text value that is not a well-formed JSON object, array, orstring is passed into JSON function, that function will usually throwan error. (Exceptions to this rule are json_valid(),json_quote(), and json_error_position().)

These routines understand allrfc-8259 JSON syntaxand also JSON5 extensions. JSON textgenerated by these routines always strictly conforms to thecanonical JSON definition and does not contain any JSON5or other extensions. The ability to read and understand JSON5 was added inversion 3.42.0 (2023-05-16).Prior versions of SQLite would only read canonical JSON.

3.2. JSONB

Beginning with version 3.45.0 (2024-01-15), SQLite allows itsinternal "parse tree" representation of JSON to be stored on disk,as a BLOB, in a format that we call "JSONB". By storing SQLite's internalbinary representation of JSON directly in the database, applicationscan bypass the overhead of parsing and rendering JSON when reading andupdating JSON values. The internal JSONB format also uses slightlyless disk space then text JSON.

Any SQL function parameter that accepts text JSON as an input will alsoaccept a BLOB in the JSONB format. The function will operate thesame in either case, except that it will run faster whenthe input is JSONB, since it does not need to run the JSON parser.

Most SQL functions that return JSON text have a corresponding functionthat returns the equivalent JSONB. The functions that return JSONin the text format begin with "json_" and functions thatreturn the binary JSONB format begin with "jsonb_".

3.2.1. The JSONB format

JSONB is a binary representation of JSON used by SQLite andis intended for internal use by SQLite only. Applicationsshould not use JSONB outside of SQLite nor try to reverse-engineer theJSONB format.

The "JSONB" name is inspired by PostgreSQL, but theon-disk format for SQLite's JSONB is not the same as PostgreSQL's.The two formats have the same name, but are not binary compatible.The PostgreSQL JSONB format claims to offer O(1)lookup of elements in objects and arrays. SQLite's JSONB format makes nosuch claim. SQLite's JSONB has O(N) time complexity formost operations in SQLite, just like text JSON. The advantage of JSONB inSQLite is that it is smaller and faster than text JSON - potentially severaltimes faster. There is space in theon-disk JSONB format to add enhancements and future versions of SQLite mightinclude options to provide O(1) lookup of elements in JSONB, but no suchcapability is currently available.

3.2.2. Handling of malformed JSONB

The JSONB that is generated by SQLite will always be well-formed. If youfollow recommended practice andtreat JSONB as an opaque BLOB, then you will not have any problems. ButJSONB is just a BLOB, so a mischievous programmer could devise BLOBsthat are similar to JSONB but that are technically malformed. Whenmisformatted JSONB is feed into JSON functions, any of the followingmight happen:

• The SQL statement might abort with a "malformed JSON" error.

• The correct answer might be returned, if the malformed parts ofthe JSONB blob do not impact the answer.

• A goofy or nonsensical answer might be returned.

The way in which SQLite handles invalid JSONB might changefrom one version of SQLite to the next. The system followsthe garbage-in/garbage-out rule: If you feed the JSON functions invalidJSONB, you get back an invalid answer. If you are in doubt about thevalidity of our JSONB, use the json_valid() function to verify it.

We do make this one promise:Malformed JSONB will never cause a memoryerror or similar problem that might lead to a vulnerability.Invalid JSONB might lead to crazy answers,or it might cause queries to abort, but it won't cause a crash.

3.3. PATH arguments

For functions that accept PATH arguments, that PATH must be well-formed orelse the function will throw an error.A well-formed PATH is a text value that begins with exactly one'$' character followed by zero or more instancesof ".objectlabel" or "[arrayindex]".

The arrayindex is usually a non-negative integer N. Inthat case, the array element selected is the N-th elementof the array, starting with zero on the left.The arrayindex can also be of the form "#-N"in which case the element selected is the N-th from theright. The last element of the array is "#-1". Think ofthe "#" characters as the "number of elements in the array". Thenthe expression "#-1" evaluates to the integer that corresponds to the last entry in the array. It is sometimes useful for the arrayindex to be just the # character, for example when appendinga value to an existing JSON array:

• json_set('[0,1,2]','$[#]','new')→ '[0,1,2,"new"]'

3.4. VALUE arguments

For functions that accept "value" arguments (also shown as"value1" and "value2"),those arguments are usually understoodto be literal strings that are quoted and become JSON string valuesin the result. Even if the input value strings look like well-formed JSON, they are still interpreted as literal strings in theresult.

However, if a value argument comes directly from the result of anotherJSON function or from the -> operator (but not the ->> operator),then the argument is understood to be actual JSON andthe complete JSON is inserted rather than a quoted string.

For example, in the following call to json_object(), the valueargument looks like a well-formed JSON array. However, because it is justordinary SQL text, it is interpreted as a literal string and added to theresult as a quoted string:

• json_object('ex','[52,3.14159]')→ '{"ex":"[52,3.14159]"}'
• json_object('ex',('[52,3.14159]'->>'$'))→ '{"ex":"[52,3.14159]"}'

But if the value argument in the outer json_object() call is theresult of another JSON function like json() or json_array(), thenthe value is understood to be actual JSON and is inserted as such:

• json_object('ex',json('[52,3.14159]'))→ '{"ex":[52,3.14159]}'
• json_object('ex',json_array(52,3.14159))→ '{"ex":[52,3.14159]}'
• json_object('ex','[52,3.14159]'->'$')→ '{"ex":[52,3.14159]}'

To be clear: "json" arguments are always interpreted as JSONregardless of where the value for that argument comes from. But"value" arguments are only interpreted as JSON if those argumentscome directly from another JSON function or the -> operator.

Within JSON value arguments interpreted as JSON strings, Unicode escapesequences are not treated as equivalent to the characters or escapedcontrol characters represented by the expressed Unicode code point.Such escape sequences are not translated or specially treated; theyare treated as plain text by SQLite's JSON functions.

3.5. Compatibility

The current implementation of this JSON library uses a recursive descentparser. In order to avoid using excess stack space, any JSON input that hasmore than 1000 levels of nesting is considered invalid. Limits on nestingdepth are allowed for compatible implementations of JSON byRFC-8259 section 9.

3.6. JSON5 Extensions

Beginning in version 3.42.0 (2023-05-16), these routines willread and interpret input JSON text that includesJSON5 extensions. However, JSON text generatedby these routines will always be strictly conforming to the canonical definition of JSON.

Here is a synopsis of JSON5 extensions (adapted from theJSON5 specification):

• Object keys may be unquoted identifiers.
• Objects may have a single trailing comma.
• Arrays may have a single trailing comma.
• Strings may be single quoted.
• Strings may span multiple lines by escaping new line characters.
• Strings may include new character escapes.
• Numbers may be hexadecimal.
• Numbers may have a leading or trailing decimal point.
• Numbers may be "Infinity", "-Infinity", and "NaN".
• Numbers may begin with an explicit plus sign.
• Single (//...) and multi-line (/*...*/) comments are allowed.
• Additional white space characters are allowed.

To convert string X from JSON5 into canonical JSON, invoke"json(X)". The output of the "json()" function will be canonicalJSON regardless of any JSON5 extensions that are present in the input.For backwards compatibility, the json_valid(X) function without a"flags" argument continuesto report false for inputs that are not canonical JSON, even if theinput is JSON5 that the function is able to understand. To determinewhether or not an input string is valid JSON5, include the 0x02 bitin the "flags" argument to json_valid: "json_valid(X,2)".

These routines understand all of JSON5, plus a little more.SQLite extends the JSON5 syntax in these two ways:

1. Strict JSON5 requires thatunquoted object keys must be ECMAScript 5.1 IdentifierNames. But largeunicode tables and lots of code is required in order to determine whether ornot a key is an ECMAScript 5.1 IdentifierName. For this reason,SQLite allows object keys to include any unicode charactersgreater than U+007f that are not whitespace characters. This relaxeddefinition of "identifier" greatly simplifies the implementation and allowsthe JSON parser to be smaller and run faster.

2. JSON5 allows floating-point infinities to be expressed as"Infinity", "-Infinity", or "+Infinity"in exactly that case - the initial "I" is capitalized and all othercharacters are lower case. SQLite also allows the abbreviation "Inf"to be used in place of "Infinity" and it allows both keywordsto appear in any combination of upper and lower case letters.Similarly,JSON5 allows "NaN" for not-a-number. SQLite extends this to also allow"QNaN" and "SNaN" in any combination of upper and lower case letters.Note that SQLite interprets NaN, QNaN, and SNaN as just an alternativespellings for "null".This extension has been added because (we are told) there exists a lotof JSON in the wild that includes these non-standard representationsfor infinity and not-a-number.

3.7. Performance Considerations

Most JSON functions do their internal processing using JSONB. So if theinput is text, they first must translate the input text into JSONB.If the input is already in the JSONB format, no translation is needed,that step can be skipped, and performance is faster.

For that reason,when an argument to one JSON function is supplied by anotherJSON function, it is usually more efficient to use the "jsonb_"variant for the function used as the argument.

• ... json_insert(A,'$.b',json(C)) ... ← Less efficient.
• ... json_insert(A,'$.b',jsonb(C)) ... ← More efficient.

The aggregate JSON SQL functions are an exception to this rule. Thosefunctions all do their processing using text instead of JSONB. So for theaggregate JSON SQL functions, it is more efficient for the argumentsto be supplied using "json_" functions than "jsonb_"functions.

• ... json_group_array(json(A))) ... ← More efficient.
• ... json_group_array(jsonb(A))) ... ← Less efficient.

3.8. The JSON BLOB Input Bug

If a JSON input is a BLOB that is not JSONB and that looks liketext JSON when cast to text, then it is accepted as text JSON.This is actually a long-standing bug in the original implementationthat the SQLite developers were unaware of. The documentation statedthat a BLOB input to a JSON function should raise an error. But in theactual implementation, the input would be accepted as longas the BLOB content was a valid JSON string in the text encoding ofthe database.

This JSON BLOB input bug was accidentally fixed when the JSON routineswere reimplemented for the 3.45.0 release (2024-01-15).That caused breakage in applications that had come to depend on the oldbehavior. (In defense of those applications: they were often lured intousing BLOBs as JSON by the readfile() SQL functionavailable in the CLI. Readfile() was used to read JSON from disk files,but readfile() returns a BLOB. And that worked for them, so why not justdo it?)

For backwards compatibility,the (formerly incorrect) legacy behavior of interpreting BLOBs as text JSONif no other interpretation worksis hereby documented and is be officially supported in version 3.45.1 (2024-01-30) and all subsequent releases.

The following sections provide additional detail on the operation ofthe various JSON functions and operators:

4.1. The json() function

The json(X) function verifies that its argument X is a validJSON string or JSONB blob and returns a minified version of that JSON stringwith all unnecessary whitespace removed. If X is not a well-formedJSON string or JSONB blob, then this routine throws an error.

If the input is JSON5 text, then it is converted into canonicalRFC-8259 text prior to being returned.

If the argument X to json(X) contains JSON objects with duplicatelabels, then it is undefined whether or not the duplicates arepreserved. The current implementation preserves duplicates.However, future enhancementsto this routine may choose to silently remove duplicates.

Example:

• json(' { "this" : "is", "a": [ "test" ] } ')→ '{"this":"is","a":["test"]}'

4.2. The jsonb() function

The jsonb(X) function returns the binary JSONB representationof the JSON provided as argument X. An error is raised if X isTEXT that does not have valid JSON syntax.

If X is a BLOB and appears to be JSONB,then this routine simply returns a copy of X.Only the outer-most element of the JSONB input is examined, however.The deep structure of the JSONB is not validated.

4.3. The json_array() function

The json_array() SQL function accepts zero or more arguments andreturns a well-formed JSON array that is composed from those arguments.If any argument to json_array() is a BLOB then an error is thrown.

An argument with SQL type TEXT is normally converted into a quoted JSON string. However, if the argument is the output from another json1function, then it is stored as JSON. This allows calls to json_array()and json_object() to be nested. The json() function can alsobe used to force strings to be recognized as JSON.

Examples:

• json_array(1,2,'3',4)→ '[1,2,"3",4]'
• json_array('[1,2]')→ '["[1,2]"]'
• json_array(json_array(1,2))→ '[[1,2]]'
• json_array(1,null,'3','[4,5]','{"six":7.7}')→ '[1,null,"3","[4,5]","{\"six\":7.7}"]'
• json_array(1,null,'3',json('[4,5]'),json('{"six":7.7}'))→ '[1,null,"3",[4,5],{"six":7.7}]'

4.4. The jsonb_array() function

The jsonb_array() SQL function works just like the json_array()function except that it returns the constructed JSON array in theSQLite's private JSONB format rather than in the standardRFC 8259 text format.

4.5. The json_array_length() function

The json_array_length(X) function returns the number of elementsin the JSON array X, or 0 if X is some kind of JSON value otherthan an array. The json_array_length(X,P) locates the array at path Pwithin X and returns the length of that array, or 0 if path P locatesan element in X that is not a JSON array, and NULL if path P does notlocate any element of X. Errors are thrown if either X is not well-formed JSON or if P is not a well-formed path.

Examples:

• json_array_length('[1,2,3,4]')→ 4
• json_array_length('[1,2,3,4]', '$')→ 4
• json_array_length('[1,2,3,4]', '$[2]')→ 0
• json_array_length('{"one":[1,2,3]}')→ 0
• json_array_length('{"one":[1,2,3]}', '$.one')→ 3
• json_array_length('{"one":[1,2,3]}', '$.two')→ NULL

4.6. The json_error_position() function

The json_error_positionf(X) function returns 0 if the input X is awell-formed JSON or JSON5 string. If the input X contains one or moresyntax errors, then this function returns the character position of thefirst syntax error. The left-most character is position 1.

If the input X is a BLOB, then this routine returns 0 if X isa well-formed JSONB blob. If the return value is positive, then itrepresents the approximate 1-based position in the BLOB of thefirst detected error.

The json_error_position() function was added withSQLite version 3.42.0 (2023-05-16).

4.7. The json_extract() function

The json_extract(X,P1,P2,...) extracts and returns one or more values from thewell-formed JSON at X. If only a single path P1 is provided, then theSQL datatype of the result is NULL for a JSON null, INTEGER or REALfor a JSON numeric value, an INTEGER zero for a JSON false value,an INTEGER one for a JSON true value, the dequoted text for a JSON string value, and a text representation for JSON object and array values.If there are multiple path arguments (P1, P2, and so forth) then thisroutine returns SQLite text which is a well-formed JSON array holdingthe various values.

Examples:

• json_extract('{"a":2,"c":[4,5,{"f":7}]}', '$')→ '{"a":2,"c":[4,5,{"f":7}]}'
• json_extract('{"a":2,"c":[4,5,{"f":7}]}', '$.c')→ '[4,5,{"f":7}]'
• json_extract('{"a":2,"c":[4,5,{"f":7}]}', '$.c[2]')→ '{"f":7}'
• json_extract('{"a":2,"c":[4,5,{"f":7}]}', '$.c[2].f')→ 7
• json_extract('{"a":2,"c":[4,5],"f":7}','$.c','$.a')→ '[[4,5],2]'
• json_extract('{"a":2,"c":[4,5],"f":7}','$.c[#-1]')→ 5
• json_extract('{"a":2,"c":[4,5,{"f":7}]}', '$.x')→ NULL
• json_extract('{"a":2,"c":[4,5,{"f":7}]}', '$.x', '$.a')→ '[null,2]'
• json_extract('{"a":"xyz"}', '$.a')→ 'xyz'
• json_extract('{"a":null}', '$.a')→ NULL

There is a subtle incompatibility between the json_extract() functionin SQLite and the json_extract() function in MySQL. The MySQL versionof json_extract() always returns JSON. The SQLite version ofjson_extract() only returns JSON if there are two or more PATH arguments(because the result is then a JSON array) or if the single PATH argumentreferences an array or object. In SQLite, if json_extract() has onlya single PATH argument and that PATH references a JSON null or a stringor a numeric value, then json_extract() returns the corresponding SQLNULL, TEXT, INTEGER, or REAL value.

The difference between MySQL json_extract() and SQLite json_extract()really only stands out when accessing individual values within the JSONthat are strings or NULLs. The following table demonstrates the difference:

4.8. The jsonb_extract() function

The jsonb_extract() function works the same as the json_extract() function,except in cases where json_extract() would normally return a textJSON array object, this routine returns the array or object in theJSONB format. For the common case where a text, numeric, null, orboolean JSON element is returned, this routine works exactly the sameas json_extract().

4.9. The -> and ->> operators

Beginning with SQLite version 3.38.0 (2022-02-22), the ->and ->> operators are available for extracting subcomponents of JSON.The SQLite implementation of -> and ->> strives to becompatible with both MySQL and PostgreSQL.The -> and ->> operators take a JSON string or JSONB blobas their left operand and a PATH expression or object fieldlabel or array index as their right operand. The -> operatorreturns a text JSON representation of the selected subcomponent orNULL if that subcomponent does not exist. The ->> operator returnsan SQL TEXT, INTEGER, REAL, or NULL value that represents the selectedsubcomponent, or NULL if the subcomponent does not exist.

Both the -> and ->> operators select the same subcomponentof the JSON to their left. The difference is that -> always returns aJSON representation of that subcomponent and the ->> operator alwaysreturns an SQL representation of that subcomponent. Thus, these operatorsare subtly different from a two-argument json_extract() function call.A call to json_extract() with two arguments will return a JSON representationof the subcomponent if and only if the subcomponent is a JSON array orobject, and will return an SQL representation of the subcomponent if thesubcomponent is a JSON null, string, or numeric value.

When the -> operator returns JSON, it always returns theRFC 8565 text representation of that JSON, not JSONB. Use thejsonb_extract() function if you need a subcomponent in theJSONB format.

The right-hand operand to the -> and ->> operators canbe a well-formed JSON path expression. This is the form used by MySQL.For compatibility with PostgreSQL,the -> and ->> operators also accept a text object label orinteger array index as their right-hand operand.If the right operand is a textlabel X, then it is interpreted as the JSON path '$.X'. If the rightoperand is an integer value N, then it is interpreted as the JSON path '$[N]'.

Examples:

• '{"a":2,"c":[4,5,{"f":7}]}' -> '$'→ '{"a":2,"c":[4,5,{"f":7}]}'
• '{"a":2,"c":[4,5,{"f":7}]}' -> '$.c'→ '[4,5,{"f":7}]'
• '{"a":2,"c":[4,5,{"f":7}]}' -> 'c'→ '[4,5,{"f":7}]'
• '{"a":2,"c":[4,5,{"f":7}]}' -> '$.c[2]'→ '{"f":7}'
• '{"a":2,"c":[4,5,{"f":7}]}' -> '$.c[2].f'→ '7'
• '{"a":2,"c":[4,5,{"f":7}]}' ->> '$.c[2].f'→ 7
• '{"a":2,"c":[4,5,{"f":7}]}' -> 'c' -> 2 ->> 'f'→ 7
• '{"a":2,"c":[4,5],"f":7}' -> '$.c[#-1]'→ '5'
• '{"a":2,"c":[4,5,{"f":7}]}' -> '$.x'→ NULL
• '[11,22,33,44]' -> 3→ '44'
• '[11,22,33,44]' ->> 3→ 44
• '{"a":"xyz"}' -> '$.a'→ '"xyz"'
• '{"a":"xyz"}' ->> '$.a'→ 'xyz'
• '{"a":null}' -> '$.a'→ 'null'
• '{"a":null}' ->> '$.a'→ NULL

4.10. The json_insert(), json_replace, and json_set() functions

The json_insert(), json_replace, and json_set() functions all takea single JSON value as their first argument followed by zero or morepairs of path and value arguments, and return a new JSON string formedby updating the input JSON by the path/value pairs. The functionsdiffer only in how they deal with creating new values and overwritingpreexisting values.

The json_insert(), json_replace(), and json_set() functions alwaystake an odd number of arguments. The first argument is always the originalJSON to be edited. Subsequent arguments occur in pairs with the firstelement of each pair being a path and the second element being the valueto insert or replace or set on that path.

Edits occur sequentially from left to right. Changes caused byprior edits can affect the path search for subsequent edits.

If the value of a path/value pair is an SQLite TEXT value, then itis normally inserted as a quoted JSON string, even if the string lookslike valid JSON. However, if the value is the result of anotherjson function (such as json() or json_array() or json_object())or if it is the result of the -> operator,then it is interpreted as JSON and is inserted as JSON retaining allof its substructure. Values that are the result of the ->> operatorare always interpreted as TEXT and are inserted as a JSON string evenif they look like valid JSON.

These routines throw an error if the first JSON argument is notwell-formed or if any PATH argument is not well-formed or if anyargument is a BLOB.

To append an element onto the end of an array, using json_insert()with an array index of "#". Examples:

• json_insert('[1,2,3,4]','$[#]',99)→ '[1,2,3,4,99]'
• json_insert('[1,[2,3],4]','$[1][#]',99)→ '[1,[2,3,99],4]'

Other examples:

• json_insert('{"a":2,"c":4}', '$.a', 99)→ '{"a":2,"c":4}'
• json_insert('{"a":2,"c":4}', '$.e', 99)→ '{"a":2,"c":4,"e":99}'
• json_replace('{"a":2,"c":4}', '$.a', 99)→ '{"a":99,"c":4}'
• json_replace('{"a":2,"c":4}', '$.e', 99)→ '{"a":2,"c":4}'
• json_set('{"a":2,"c":4}', '$.a', 99)→ '{"a":99,"c":4}'
• json_set('{"a":2,"c":4}', '$.e', 99)→ '{"a":2,"c":4,"e":99}'
• json_set('{"a":2,"c":4}', '$.c', '[97,96]')→ '{"a":2,"c":"[97,96]"}'
• json_set('{"a":2,"c":4}', '$.c', json('[97,96]'))→ '{"a":2,"c":[97,96]}'
• json_set('{"a":2,"c":4}', '$.c', json_array(97,96))→ '{"a":2,"c":[97,96]}'

4.11. The jsonb_insert(), jsonb_replace, and jsonb_set() functions

The jsonb_insert(), jsonb_replace(), and jsonb_set() functions work thesame as json_insert(), json_replace(), and json_set(), respectively,except that "jsonb_" versions return their result in the binaryJSONB format.

4.12. The json_object() function

The json_object() SQL function accepts zero or more pairs of argumentsand returns a well-formed JSON object that is composed from those arguments.The first argument of each pair is the label and the second argument ofeach pair is the value.If any argument to json_object() is a BLOB then an error is thrown.

The json_object() function currently allows duplicate labels withoutcomplaint, though this might change in a future enhancement.

An argument with SQL type TEXT it is normally converted into a quoted JSON string even if the input text is well-formed JSON. However, if the argument is the direct result from another JSONfunction or the -> operator (but not the ->> operator), then it is treated as JSON and all of its JSON type informationand substructure is preserved. This allows calls to json_object()and json_array() to be nested. The json() function can alsobe used to force strings to be recognized as JSON.

Examples:

• json_object('a',2,'c',4)→ '{"a":2,"c":4}'
• json_object('a',2,'c','{e:5}')→ '{"a":2,"c":"{e:5}"}'
• json_object('a',2,'c',json_object('e',5))→ '{"a":2,"c":{"e":5}}'

4.13. The jsonb_object() function

The jsonb_object() function works just like the json_object() functionexcept that the generated object is returned in the binary JSONB format.

4.14. The json_patch() function

The json_patch(T,P) SQL function runs theRFC-7396 MergePatch algorithmto apply patch P against input T. The patched copy of T is returned.

MergePatch can add, modify, or delete elements of a JSON Object,and so for JSON Objects, the json_patch() routine is a generalizedreplacement for json_set() and json_remove(). However, MergePatchtreats JSON Array objects as atomic. MergePatch cannot append to anArray nor modify individual elements of an Array. It can only insert,replace, or delete the whole Array as a single unit. Hence, json_patch()is not as useful when dealing with JSON that includes Arrays,especially Arrays with lots of substructure.

Examples:

• json_patch('{"a":1,"b":2}','{"c":3,"d":4}')→ '{"a":1,"b":2,"c":3,"d":4}'
• json_patch('{"a":[1,2],"b":2}','{"a":9}')→ '{"a":9,"b":2}'
• json_patch('{"a":[1,2],"b":2}','{"a":null}')→ '{"b":2}'
• json_patch('{"a":1,"b":2}','{"a":9,"b":null,"c":8}')→ '{"a":9,"c":8}'
• json_patch('{"a":{"x":1,"y":2},"b":3}','{"a":{"y":9},"c":8}')→ '{"a":{"x":1,"y":9},"b":3,"c":8}'

4.15. The jsonb_patch() function

The jsonb_patch() function works just like the json_patch() functionexcept that the patched JSON is returned in the binary JSONB format.

4.16. The json_remove() function

The json_remove(X,P,...) function takes a single JSON value as itsfirst argument followed by zero or more path arguments.The json_remove(X,P,...) function returnsa copy of the X parameter with all the elements identified by path arguments removed. Paths that select elementsnot found in X are silently ignored.

Removals occurs sequentially from left to right. Changes caused byprior removals can affect the path search for subsequent arguments.

If the json_remove(X) function is called with no path arguments,then it returns the input X reformatted, with excess whitespaceremoved.

The json_remove() function throws an error if the first argumentis not well-formed JSON or if any later argument is not a well-formedpath.

Examples:

• json_remove('[0,1,2,3,4]','$[2]')→ '[0,1,3,4]'
• json_remove('[0,1,2,3,4]','$[2]','$[0]')→ '[1,3,4]'
• json_remove('[0,1,2,3,4]','$[0]','$[2]')→ '[1,2,4]'
• json_remove('[0,1,2,3,4]','$[#-1]','$[0]')→ '[1,2,3]'
• json_remove('{"x":25,"y":42}')→ '{"x":25,"y":42}'
• json_remove('{"x":25,"y":42}','$.z')→ '{"x":25,"y":42}'
• json_remove('{"x":25,"y":42}','$.y')→ '{"x":25}'
• json_remove('{"x":25,"y":42}','$')→ NULL

4.17. The jsonb_remove() function

The jsonb_remove() function works just like the json_remove() functionexcept that the edited JSON result is returned in the binary JSONB format.

4.18. The json_type() function

The json_type(X) function returns the "type" of the outermost elementof X. The json_type(X,P) function returns the "type" of the elementin X that is selected by path P. The "type" returned by json_type() isone of the following SQL text values:'null', 'true', 'false', 'integer', 'real', 'text', 'array', or 'object'.If the path P in json_type(X,P) selects an element that does not existin X, then this function returns NULL.

The json_type() function throws an error if its first argument isnot well-formed JSON or JSONB or if its second argument is not a well-formedJSON path.

Examples:

• json_type('{"a":[2,3.5,true,false,null,"x"]}')→ 'object'
• json_type('{"a":[2,3.5,true,false,null,"x"]}','$')→ 'object'
• json_type('{"a":[2,3.5,true,false,null,"x"]}','$.a')→ 'array'
• json_type('{"a":[2,3.5,true,false,null,"x"]}','$.a[0]')→ 'integer'
• json_type('{"a":[2,3.5,true,false,null,"x"]}','$.a[1]')→ 'real'
• json_type('{"a":[2,3.5,true,false,null,"x"]}','$.a[2]')→ 'true'
• json_type('{"a":[2,3.5,true,false,null,"x"]}','$.a[3]')→ 'false'
• json_type('{"a":[2,3.5,true,false,null,"x"]}','$.a[4]')→ 'null'
• json_type('{"a":[2,3.5,true,false,null,"x"]}','$.a[5]')→ 'text'
• json_type('{"a":[2,3.5,true,false,null,"x"]}','$.a[6]')→ NULL

4.19. The json_valid() function

The json_valid(X,Y) function return 1 if the argument X is well-formedJSON, or returns 0 if X is not well-formed. The Y parameter is an integerbitmask that defines what is meant by "well-formed". The following bitsof Y are currently defined:

• 0x01 →The input is text that strictly complies with canonical RFC-8259 JSON,without any extensions.
• 0x02 →The input is text that is JSON with JSON5 extensions described above.
• 0x04 →The input is a BLOB that superficially appears to be JSONB.
• 0x08 →The input is a BLOB that strictly conforms to the internal JSONB format.

By combining bits, the following useful values of Y can be derived:

• 1 → X is RFC-8259 JSON text
• 2 → X is JSON5 text
• 4 → X is probably JSONB
• 5 → X is RFC-8259 JSON text or JSONB
• 6 → X is JSON5 text or JSONB ← This is probably the value you want
• 8 → X is strictly conforming JSONB
• 9 → X is RFC-8259 or strictly conforming JSONB
• 10 → X is JSON5 or strictly conforming JSONB

The Y parameter is optional. If omitted, it defaults to 1, which meansthat the default behavior is to return true only if the input X isstrictly conforming RFC-8259 JSON text without any extensions. Thismakes the one-argument version of json_valid() compatible with olderversions of SQLite, prior to the addition of support forJSON5 and JSONB.

The difference between 0x04 and 0x08 bits in the Y parameter is that0x04 only examines the outer wrapper of the BLOB to see if it superficiallylooks like JSONB. This is sufficient for must purposes and is very fast.The 0x08 bit does a thorough examination of all internal details of the BLOB.The 0x08 bit takes time that is linear in the size of the X input and is muchslower. The 0x04 bit is recommended for most purposes.

If you just want to know if a value is a plausible input to one ofthe other JSON functions, a Y value of 6 is probably what you want to use.

Any Y value less than 1 or greater than 15 raises an error, for thelatest version of json_valid(). However, future versions of json_valid()might be enhanced to accept flag values outside of this range, having newmeanings that we have not yet thought of.

If either X or Y inputs to json_valid() are NULL, then the functionreturns NULL.

Examples:

• json_valid('{"x":35}')→ 1
• json_valid('{x:35}')→ 0
• json_valid('{x:35}',6)→ 1
• json_valid('{"x":35')→ 0
• json_valid(NULL)→ NULL

4.20. The json_quote() function

The json_quote(X) function converts the SQL value X (a number or astring) into its corresponding JSON representation. If X is a JSON valuereturned by another JSON function, then this function is a no-op.

Examples:

• json_quote(3.14159)→ 3.14159
• json_quote('verdant')→ '"verdant"'
• json_quote('[1]')→ '"[1]"'
• json_quote(json('[1]'))→ '[1]'
• json_quote('[1,')→ '"[1,"'

4.21. Array and object aggregate functions

The json_group_array(X) function is anaggregate SQL function that returns a JSON arraycomprised of all X values in the aggregation.Similarly, the json_group_object(NAME,VALUE) function returns a JSON objectcomprised of all NAME/VALUE pairs in the aggregation.The "jsonb_" variants are the same except that they return theirresult in the binary JSONB format.

4.22. The json_each() and json_tree() table-valued functions

The json_each(X) and json_tree(X) table-valued functions walk theJSON value provided as their first argument and return one row for eachelement. The json_each(X) function only walks the immediate childrenof the top-level array or object,or just the top-level element itself if the top-levelelement is a primitive value.The json_tree(X) function recursively walks through theJSON substructure starting with the top-level element.

The json_each(X,P) and json_tree(X,P) functions work just liketheir one-argument counterparts except that they treat the elementidentified by path P as the top-level element.

The schema for the table returned by json_each() and json_tree() isas follows:

CREATE TABLE json_tree( key ANY, -- key for current element relative to its parent value ANY, -- value for the current element type TEXT, -- 'object','array','string','integer', etc. atom ANY, -- value for primitive types, null for array & object id INTEGER, -- integer ID for this element parent INTEGER, -- integer ID for the parent of this element fullkey TEXT, -- full path describing the current element path TEXT, -- path to the container of the current row json JSON HIDDEN, -- 1st input parameter: the raw JSON root TEXT HIDDEN -- 2nd input parameter: the PATH at which to start);

The "key" column is the integer array index for elements of a JSON array and the text label for elements of a JSON object. The key column isNULL in all other cases.

The "atom" column is the SQL value corresponding to primitive elements - elements other than JSON arrays and objects. The "atom" column is NULLfor a JSON array or object. The "value" column is the same as the"atom" column for primitive JSON elements but takes on the text JSON valuefor arrays and objects.

The "type" column is an SQL text value taken from ('null', 'true', 'false','integer', 'real', 'text', 'array', 'object') according to the type ofthe current JSON element.

The "id" column is an integer that identifies a specific JSON elementwithin the complete JSON string. The "id" integer is an internal housekeepingnumber, the computation of which might change in future releases. Theonly guarantee is that the "id" column will be different for every row.

The "parent" column is always NULL for json_each().For json_tree(),the "parent" column is the "id" integer for the parent of the currentelement, or NULL for the top-level JSON element or the element identifiedby the root path in the second argument.

The "fullkey" column is a text path that uniquely identifies the currentrow element within the original JSON string. The complete key to thetrue top-level element is returned even if an alternative starting pointis provided by the "root" argument.

The "path" column is the path to the array or object container that holds the current row, or the path to the current row in the case where the iteration starts on a primitive type and thus only provides a singlerow of output.

4.22.1. Examples using json_each() and json_tree()

Suppose the table "CREATE TABLE user(name,phone)" stores zero ormore phone numbers as a JSON array object in the user.phone field.To find all users who have any phone number with a 704 area code:

SELECT DISTINCT user.name FROM user, json_each(user.phone) WHERE json_each.value LIKE '704-%';

Now suppose the user.phone field contains plain text if the userhas only a single phone number and a JSON array if the user has multiplephone numbers. The same question is posed: "Which users have a phone numberin the 704 area code?" But now the json_each() function can only be calledfor those users that have two or more phone numbers since json_each()requires well-formed JSON as its first argument:

SELECT name FROM user WHERE phone LIKE '704-%'UNIONSELECT user.name FROM user, json_each(user.phone) WHERE json_valid(user.phone) AND json_each.value LIKE '704-%';

Consider a different database with "CREATE TABLE big(json JSON)".To see a complete line-by-line decomposition of the data:

SELECT big.rowid, fullkey, value FROM big, json_tree(big.json) WHERE json_tree.type NOT IN ('object','array');

In the previous, the "type NOT IN ('object','array')" term of theWHERE clause suppresses containers and only lets through leaf elements.The same effect could be achieved this way:

SELECT big.rowid, fullkey, atom FROM big, json_tree(big.json) WHERE atom IS NOT NULL;

Suppose each entry in the BIG table is a JSON object with a '$.id' field that is a unique identifierand a '$.partlist' field that can be a deeply nested object.You want to find the id of every entry that contains oneor more references to uuid '6fa5181e-5721-11e5-a04e-57f3d7b32808' anywherein its '$.partlist'.

SELECT DISTINCT json_extract(big.json,'$.id') FROM big, json_tree(big.json, '$.partlist') WHERE json_tree.key='uuid' AND json_tree.value='6fa5181e-5721-11e5-a04e-57f3d7b32808';

This page last modified on 2024-03-14 14:17:16 UTC