postgresql/doc/src/sgml/json.sgml

<!-- doc/src/sgml/json.sgml -->

<sect1 id="datatype-json">
 <title><acronym>JSON</acronym> Types</title>

 <indexterm zone="datatype-json">
  <primary>JSON</primary>
 </indexterm>

 <indexterm zone="datatype-json">
  <primary>JSONB</primary>
 </indexterm>

 <para>
  JSON data types are for storing JSON (JavaScript Object Notation)
  data, as specified in <ulink url="https://datatracker.ietf.org/doc/html/rfc7159">RFC
  7159</ulink>. Such data can also be stored as <type>text</type>, but
  the JSON data types have the advantage of enforcing that each
  stored value is valid according to the JSON rules.  There are also
  assorted JSON-specific functions and operators available for data stored
  in these data types; see <xref linkend="functions-json"/>.
 </para>

 <para>
  <productname>PostgreSQL</productname> offers two types for storing JSON
  data: <type>json</type> and <type>jsonb</type>. To implement efficient query
  mechanisms for these data types, <productname>PostgreSQL</productname>
  also provides the <type>jsonpath</type> data type described in
  <xref linkend="datatype-jsonpath"/>.
 </para>

 <para>
  The <type>json</type> and <type>jsonb</type> data types
  accept <emphasis>almost</emphasis> identical sets of values as
  input.  The major practical difference is one of efficiency.  The
  <type>json</type> data type stores an exact copy of the input text,
  which processing functions must reparse on each execution; while
  <type>jsonb</type> data is stored in a decomposed binary format that
  makes it slightly slower to input due to added conversion
  overhead, but significantly faster to process, since no reparsing
  is needed.  <type>jsonb</type> also supports indexing, which can be a
  significant advantage.
 </para>

 <para>
  Because the <type>json</type> type stores an exact copy of the input text, it
  will preserve semantically-insignificant white space between tokens, as
  well as the order of keys within JSON objects. Also, if a JSON object
  within the value contains the same key more than once, all the key/value
  pairs are kept.  (The processing functions consider the last value as the
  operative one.)  By contrast, <type>jsonb</type> does not preserve white
  space, does not preserve the order of object keys, and does not keep
  duplicate object keys.  If duplicate keys are specified in the input,
  only the last value is kept.
 </para>

 <para>
  In general, most applications should prefer to store JSON data as
  <type>jsonb</type>, unless there are quite specialized needs, such as
  legacy assumptions about ordering of object keys.
 </para>

 <para>
  <acronym>RFC</acronym> 7159 specifies that JSON strings should be encoded in UTF8.
  It is therefore not possible for the JSON
  types to conform rigidly to the JSON specification unless the database
  encoding is UTF8. Attempts to directly include characters that
  cannot be represented in the database encoding will fail; conversely,
  characters that can be represented in the database encoding but not
  in UTF8 will be allowed.
 </para>

 <para>
  <acronym>RFC</acronym> 7159 permits JSON strings to contain Unicode escape sequences
  denoted by <literal>\u<replaceable>XXXX</replaceable></literal>.  In the input
  function for the <type>json</type> type, Unicode escapes are allowed
  regardless of the database encoding, and are checked only for syntactic
  correctness (that is, that four hex digits follow <literal>\u</literal>).
  However, the input function for <type>jsonb</type> is stricter: it disallows
  Unicode escapes for characters that cannot be represented in the database
  encoding.  The <type>jsonb</type> type also
  rejects <literal>\u0000</literal> (because that cannot be represented in
  <productname>PostgreSQL</productname>'s <type>text</type> type), and it insists
  that any use of Unicode surrogate pairs to designate characters outside
  the Unicode Basic Multilingual Plane be correct.  Valid Unicode escapes
  are converted to the equivalent single character for storage;
  this includes folding surrogate pairs into a single character.
 </para>

 <note>
  <para>
   Many of the JSON processing functions described
   in <xref linkend="functions-json"/> will convert Unicode escapes to
   regular characters, and will therefore throw the same types of errors
   just described even if their input is of type <type>json</type>
   not <type>jsonb</type>. The fact that the <type>json</type> input function does
   not make these checks may be considered a historical artifact, although
   it does allow for simple storage (without processing) of JSON Unicode
   escapes in a database encoding that does not support the represented
   characters.
  </para>
 </note>

 <para>
  When converting textual JSON input into <type>jsonb</type>, the primitive
  types described by <acronym>RFC</acronym> 7159 are effectively mapped onto
  native <productname>PostgreSQL</productname> types, as shown
  in <xref linkend="json-type-mapping-table"/>.
  Therefore, there are some minor additional constraints on what
  constitutes valid <type>jsonb</type> data that do not apply to
  the <type>json</type> type, nor to JSON in the abstract, corresponding
  to limits on what can be represented by the underlying data type.
  Notably, <type>jsonb</type> will reject numbers that are outside the
  range of the <productname>PostgreSQL</productname> <type>numeric</type> data
  type, while <type>json</type> will not.  Such implementation-defined
  restrictions are permitted by <acronym>RFC</acronym> 7159.  However, in
  practice such problems are far more likely to occur in other
  implementations, as it is common to represent JSON's <type>number</type>
  primitive type as IEEE 754 double precision floating point
  (which <acronym>RFC</acronym> 7159 explicitly anticipates and allows for).
  When using JSON as an interchange format with such systems, the danger
  of losing numeric precision compared to data originally stored
  by <productname>PostgreSQL</productname> should be considered.
 </para>

 <para>
  Conversely, as noted in the table there are some minor restrictions on
  the input format of JSON primitive types that do not apply to
  the corresponding <productname>PostgreSQL</productname> types.
 </para>

  <table id="json-type-mapping-table">
     <title>JSON Primitive Types and Corresponding <productname>PostgreSQL</productname> Types</title>
     <tgroup cols="3">
      <colspec colname="col1" colwidth="1*"/>
      <colspec colname="col2" colwidth="1*"/>
      <colspec colname="col3" colwidth="2*"/>
      <thead>
       <row>
        <entry>JSON primitive type</entry>
        <entry><productname>PostgreSQL</productname> type</entry>
        <entry>Notes</entry>
       </row>
      </thead>
      <tbody>
       <row>
        <entry><type>string</type></entry>
        <entry><type>text</type></entry>
        <entry><literal>\u0000</literal> is disallowed, as are Unicode escapes
         representing characters not available in the database encoding</entry>
       </row>
       <row>
        <entry><type>number</type></entry>
        <entry><type>numeric</type></entry>
        <entry><literal>NaN</literal> and <literal>infinity</literal> values are disallowed</entry>
       </row>
       <row>
        <entry><type>boolean</type></entry>
        <entry><type>boolean</type></entry>
        <entry>Only lowercase <literal>true</literal> and <literal>false</literal> spellings are accepted</entry>
       </row>
       <row>
        <entry><type>null</type></entry>
        <entry>(none)</entry>
        <entry>SQL <literal>NULL</literal> is a different concept</entry>
       </row>
      </tbody>
     </tgroup>
   </table>

 <sect2 id="json-keys-elements">
  <title>JSON Input and Output Syntax</title>
  <para>
   The input/output syntax for the JSON data types is as specified in
   <acronym>RFC</acronym> 7159.
  </para>
  <para>
   The following are all valid <type>json</type> (or <type>jsonb</type>) expressions:
<programlisting>
-- Simple scalar/primitive value
-- Primitive values can be numbers, quoted strings, true, false, or null
SELECT '5'::json;

-- Array of zero or more elements (elements need not be of same type)
SELECT '[1, 2, "foo", null]'::json;

-- Object containing pairs of keys and values
-- Note that object keys must always be quoted strings
SELECT '{"bar": "baz", "balance": 7.77, "active": false}'::json;

-- Arrays and objects can be nested arbitrarily
SELECT '{"foo": [true, "bar"], "tags": {"a": 1, "b": null}}'::json;
</programlisting>
  </para>

  <para>
   As previously stated, when a JSON value is input and then printed without
   any additional processing, <type>json</type> outputs the same text that was
   input, while <type>jsonb</type> does not preserve semantically-insignificant
   details such as whitespace.  For example, note the differences here:
<programlisting>
SELECT '{"bar": "baz", "balance": 7.77, "active":false}'::json;
                      json
-------------------------------------------------
 {"bar": "baz", "balance": 7.77, "active":false}
(1 row)

SELECT '{"bar": "baz", "balance": 7.77, "active":false}'::jsonb;
                      jsonb
--------------------------------------------------
 {"bar": "baz", "active": false, "balance": 7.77}
(1 row)
</programlisting>
   One semantically-insignificant detail worth noting is that
   in <type>jsonb</type>, numbers will be printed according to the behavior of the
   underlying <type>numeric</type> type.  In practice this means that numbers
   entered with <literal>E</literal> notation will be printed without it, for
   example:
<programlisting>
SELECT '{"reading": 1.230e-5}'::json, '{"reading": 1.230e-5}'::jsonb;
         json          |          jsonb
-----------------------+-------------------------
 {"reading": 1.230e-5} | {"reading": 0.00001230}
(1 row)
</programlisting>
   However, <type>jsonb</type> will preserve trailing fractional zeroes, as seen
   in this example, even though those are semantically insignificant for
   purposes such as equality checks.
  </para>

  <para>
    For the list of built-in functions and operators available for
    constructing and processing JSON values, see <xref linkend="functions-json"/>.
  </para>
 </sect2>

 <sect2 id="json-doc-design">
  <title>Designing JSON Documents</title>
  <para>
   Representing data as JSON can be considerably more flexible than
   the traditional relational data model, which is compelling in
   environments where requirements are fluid.  It is quite possible
   for both approaches to co-exist and complement each other within
   the same application.  However, even for applications where maximal
   flexibility is desired, it is still recommended that JSON documents
   have a somewhat fixed structure.  The structure is typically
   unenforced (though enforcing some business rules declaratively is
   possible), but having a predictable structure makes it easier to write
   queries that usefully summarize a set of <quote>documents</quote> (datums)
   in a table.
  </para>
  <para>
   JSON data is subject to the same concurrency-control
   considerations as any other data type when stored in a table.
   Although storing large documents is practicable, keep in mind that
   any update acquires a row-level lock on the whole row.
   Consider limiting JSON documents to a
   manageable size in order to decrease lock contention among updating
   transactions.  Ideally, JSON documents should each
   represent an atomic datum that business rules dictate cannot
   reasonably be further subdivided into smaller datums that
   could be modified independently.
  </para>
 </sect2>

 <sect2 id="json-containment">
  <title><type>jsonb</type> Containment and Existence</title>
  <indexterm>
    <primary>jsonb</primary>
    <secondary>containment</secondary>
  </indexterm>
  <indexterm>
    <primary>jsonb</primary>
    <secondary>existence</secondary>
  </indexterm>
  <para>
    Testing <firstterm>containment</firstterm> is an important capability of
    <type>jsonb</type>.  There is no parallel set of facilities for the
    <type>json</type> type.  Containment tests whether
    one <type>jsonb</type> document has contained within it another one.
    These examples return true except as noted:
  </para>
<programlisting>
-- Simple scalar/primitive values contain only the identical value:
SELECT '"foo"'::jsonb @&gt; '"foo"'::jsonb;

-- The array on the right side is contained within the one on the left:
SELECT '[1, 2, 3]'::jsonb @&gt; '[1, 3]'::jsonb;

-- Order of array elements is not significant, so this is also true:
SELECT '[1, 2, 3]'::jsonb @&gt; '[3, 1]'::jsonb;

-- Duplicate array elements don't matter either:
SELECT '[1, 2, 3]'::jsonb @&gt; '[1, 2, 2]'::jsonb;

-- The object with a single pair on the right side is contained
-- within the object on the left side:
SELECT '{"product": "PostgreSQL", "version": 9.4, "jsonb": true}'::jsonb @&gt; '{"version": 9.4}'::jsonb;

-- The array on the right side is <emphasis>not</emphasis> considered contained within the
-- array on the left, even though a similar array is nested within it:
SELECT '[1, 2, [1, 3]]'::jsonb @&gt; '[1, 3]'::jsonb;  -- yields false

-- But with a layer of nesting, it is contained:
SELECT '[1, 2, [1, 3]]'::jsonb @&gt; '[[1, 3]]'::jsonb;

-- Similarly, containment is not reported here:
SELECT '{"foo": {"bar": "baz"}}'::jsonb @&gt; '{"bar": "baz"}'::jsonb;  -- yields false

-- A top-level key and an empty object is contained:
SELECT '{"foo": {"bar": "baz"}}'::jsonb @&gt; '{"foo": {}}'::jsonb;
</programlisting>

  <para>
   The general principle is that the contained object must match the
   containing object as to structure and data contents, possibly after
   discarding some non-matching array elements or object key/value pairs
   from the containing object.
   But remember that the order of array elements is not significant when
   doing a containment match, and duplicate array elements are effectively
   considered only once.
  </para>

  <para>
   As a special exception to the general principle that the structures
   must match, an array may contain a primitive value:
  </para>
<programlisting>
-- This array contains the primitive string value:
SELECT '["foo", "bar"]'::jsonb @&gt; '"bar"'::jsonb;

-- This exception is not reciprocal -- non-containment is reported here:
SELECT '"bar"'::jsonb @&gt; '["bar"]'::jsonb;  -- yields false
</programlisting>

  <para>
    <type>jsonb</type> also has an <firstterm>existence</firstterm> operator, which is
    a variation on the theme of containment: it tests whether a string
    (given as a <type>text</type> value) appears as an object key or array
    element at the top level of the <type>jsonb</type> value.
    These examples return true except as noted:
  </para>
<programlisting>
-- String exists as array element:
SELECT '["foo", "bar", "baz"]'::jsonb ? 'bar';

-- String exists as object key:
SELECT '{"foo": "bar"}'::jsonb ? 'foo';

-- Object values are not considered:
SELECT '{"foo": "bar"}'::jsonb ? 'bar';  -- yields false

-- As with containment, existence must match at the top level:
SELECT '{"foo": {"bar": "baz"}}'::jsonb ? 'bar'; -- yields false

-- A string is considered to exist if it matches a primitive JSON string:
SELECT '"foo"'::jsonb ? 'foo';
</programlisting>

  <para>
    JSON objects are better suited than arrays for testing containment or
    existence when there are many keys or elements involved, because
    unlike arrays they are internally optimized for searching, and do not
    need to be searched linearly.
  </para>

  <tip>
   <para>
    Because JSON containment is nested, an appropriate query can skip
    explicit selection of sub-objects.  As an example, suppose that we have
    a <structfield>doc</structfield> column containing objects at the top level, with
    most objects containing <literal>tags</literal> fields that contain arrays of
    sub-objects.  This query finds entries in which sub-objects containing
    both <literal>"term":"paris"</literal> and <literal>"term":"food"</literal> appear,
    while ignoring any such keys outside the <literal>tags</literal> array:
<programlisting>
SELECT doc-&gt;'site_name' FROM websites
  WHERE doc @&gt; '{"tags":[{"term":"paris"}, {"term":"food"}]}';
</programlisting>
    One could accomplish the same thing with, say,
<programlisting>
SELECT doc-&gt;'site_name' FROM websites
  WHERE doc-&gt;'tags' @&gt; '[{"term":"paris"}, {"term":"food"}]';
</programlisting>
    but that approach is less flexible, and often less efficient as well.
   </para>

   <para>
    On the other hand, the JSON existence operator is not nested: it will
    only look for the specified key or array element at top level of the
    JSON value.
   </para>
  </tip>

  <para>
    The various containment and existence operators, along with all other
    JSON operators and functions are documented
    in <xref linkend="functions-json"/>.
  </para>
 </sect2>

 <sect2 id="json-indexing">
  <title><type>jsonb</type> Indexing</title>
  <indexterm>
    <primary>jsonb</primary>
    <secondary>indexes on</secondary>
  </indexterm>

  <para>
    GIN indexes can be used to efficiently search for
    keys or key/value pairs occurring within a large number of
    <type>jsonb</type> documents (datums).
    Two GIN <quote>operator classes</quote> are provided, offering different
    performance and flexibility trade-offs.
  </para>
  <para>
    The default GIN operator class for <type>jsonb</type> supports queries with
    the key-exists operators <literal>?</literal>, <literal>?|</literal>
    and <literal>?&amp;</literal>, the containment operator
    <literal>@&gt;</literal>, and the <type>jsonpath</type> match
    operators <literal>@?</literal> and <literal>@@</literal>.
    (For details of the semantics that these operators
    implement, see <xref linkend="functions-jsonb-op-table"/>.)
    An example of creating an index with this operator class is:
<programlisting>
CREATE INDEX idxgin ON api USING GIN (jdoc);
</programlisting>
    The non-default GIN operator class <literal>jsonb_path_ops</literal>
    does not support the key-exists operators, but it does support
    <literal>@&gt;</literal>, <literal>@?</literal> and <literal>@@</literal>.
    An example of creating an index with this operator class is:
<programlisting>
CREATE INDEX idxginp ON api USING GIN (jdoc jsonb_path_ops);
</programlisting>
  </para>

  <para>
    Consider the example of a table that stores JSON documents
    retrieved from a third-party web service, with a documented schema
    definition.  A typical document is:
<programlisting>
{
    "guid": "9c36adc1-7fb5-4d5b-83b4-90356a46061a",
    "name": "Angela Barton",
    "is_active": true,
    "company": "Magnafone",
    "address": "178 Howard Place, Gulf, Washington, 702",
    "registered": "2009-11-07T08:53:22 +08:00",
    "latitude": 19.793713,
    "longitude": 86.513373,
    "tags": [
        "enim",
        "aliquip",
        "qui"
    ]
}
</programlisting>
    We store these documents in a table named <structname>api</structname>,
    in a <type>jsonb</type> column named <structfield>jdoc</structfield>.
    If a GIN index is created on this column,
    queries like the following can make use of the index:
<programlisting>
-- Find documents in which the key "company" has value "Magnafone"
SELECT jdoc-&gt;'guid', jdoc-&gt;'name' FROM api WHERE jdoc @&gt; '{"company": "Magnafone"}';
</programlisting>
    However, the index could not be used for queries like the
    following, because though the operator <literal>?</literal> is indexable,
    it is not applied directly to the indexed column <structfield>jdoc</structfield>:
<programlisting>
-- Find documents in which the key "tags" contains key or array element "qui"
SELECT jdoc-&gt;'guid', jdoc-&gt;'name' FROM api WHERE jdoc -&gt; 'tags' ? 'qui';
</programlisting>
    Still, with appropriate use of expression indexes, the above
    query can use an index.  If querying for particular items within
    the <literal>"tags"</literal> key is common, defining an index like this
    may be worthwhile:
<programlisting>
CREATE INDEX idxgintags ON api USING GIN ((jdoc -&gt; 'tags'));
</programlisting>
    Now, the <literal>WHERE</literal> clause <literal>jdoc -&gt; 'tags' ? 'qui'</literal>
    will be recognized as an application of the indexable
    operator <literal>?</literal> to the indexed
    expression <literal>jdoc -&gt; 'tags'</literal>.
    (More information on expression indexes can be found in <xref
    linkend="indexes-expressional"/>.)
  </para>

  <para>
    Another approach to querying is to exploit containment, for example:
<programlisting>
-- Find documents in which the key "tags" contains array element "qui"
SELECT jdoc-&gt;'guid', jdoc-&gt;'name' FROM api WHERE jdoc @&gt; '{"tags": ["qui"]}';
</programlisting>
    A simple GIN index on the <structfield>jdoc</structfield> column can support this
    query.  But note that such an index will store copies of every key and
    value in the <structfield>jdoc</structfield> column, whereas the expression index
    of the previous example stores only data found under
    the <literal>tags</literal> key.  While the simple-index approach is far more
    flexible (since it supports queries about any key), targeted expression
    indexes are likely to be smaller and faster to search than a simple
    index.
  </para>

  <para>
    GIN indexes also support the <literal>@?</literal>
    and <literal>@@</literal> operators, which
    perform <type>jsonpath</type> matching.  Examples are
<programlisting>
SELECT jdoc->'guid', jdoc->'name' FROM api WHERE jdoc @? '$.tags[*] ? (@ == "qui")';
</programlisting>
<programlisting>
SELECT jdoc->'guid', jdoc->'name' FROM api WHERE jdoc @@ '$.tags[*] == "qui"';
</programlisting>
    For these operators, a GIN index extracts clauses of the form
    <literal><replaceable>accessors_chain</replaceable>
    == <replaceable>constant</replaceable></literal> out of
    the <type>jsonpath</type> pattern, and does the index search based on
    the keys and values mentioned in these clauses.  The accessors chain
    may include <literal>.<replaceable>key</replaceable></literal>,
    <literal>[*]</literal>,
    and <literal>[<replaceable>index</replaceable>]</literal> accessors.
    The <literal>jsonb_ops</literal> operator class also
    supports <literal>.*</literal> and <literal>.**</literal> accessors,
    but the <literal>jsonb_path_ops</literal> operator class does not.
  </para>

  <para>
    Although the <literal>jsonb_path_ops</literal> operator class supports
    only queries with the <literal>@&gt;</literal>, <literal>@?</literal>
    and <literal>@@</literal> operators, it has notable
    performance advantages over the default operator
    class <literal>jsonb_ops</literal>.  A <literal>jsonb_path_ops</literal>
    index is usually much smaller than a <literal>jsonb_ops</literal>
    index over the same data, and the specificity of searches is better,
    particularly when queries contain keys that appear frequently in the
    data.  Therefore search operations typically perform better
    than with the default operator class.
  </para>

  <para>
    The technical difference between a <literal>jsonb_ops</literal>
    and a <literal>jsonb_path_ops</literal> GIN index is that the former
    creates independent index items for each key and value in the data,
    while the latter creates index items only for each value in the
    data.
    <footnote>
     <para>
      For this purpose, the term <quote>value</quote> includes array elements,
      though JSON terminology sometimes considers array elements distinct
      from values within objects.
     </para>
    </footnote>
    Basically, each <literal>jsonb_path_ops</literal> index item is
    a hash of the value and the key(s) leading to it; for example to index
    <literal>{"foo": {"bar": "baz"}}</literal>, a single index item would
    be created incorporating all three of <literal>foo</literal>, <literal>bar</literal>,
    and <literal>baz</literal> into the hash value.  Thus a containment query
    looking for this structure would result in an extremely specific index
    search; but there is no way at all to find out whether <literal>foo</literal>
    appears as a key.  On the other hand, a <literal>jsonb_ops</literal>
    index would create three index items representing <literal>foo</literal>,
    <literal>bar</literal>, and <literal>baz</literal> separately; then to do the
    containment query, it would look for rows containing all three of
    these items.  While GIN indexes can perform such an AND search fairly
    efficiently, it will still be less specific and slower than the
    equivalent <literal>jsonb_path_ops</literal> search, especially if
    there are a very large number of rows containing any single one of the
    three index items.
  </para>

  <para>
    A disadvantage of the <literal>jsonb_path_ops</literal> approach is
    that it produces no index entries for JSON structures not containing
    any values, such as <literal>{"a": {}}</literal>.  If a search for
    documents containing such a structure is requested, it will require a
    full-index scan, which is quite slow.  <literal>jsonb_path_ops</literal> is
    therefore ill-suited for applications that often perform such searches.
  </para>

  <para>
    <type>jsonb</type> also supports <literal>btree</literal> and <literal>hash</literal>
    indexes.  These are usually useful only if it's important to check
    equality of complete JSON documents.
    The <literal>btree</literal> ordering for <type>jsonb</type> datums is seldom
    of great interest, but for completeness it is:
<synopsis>
<replaceable>Object</replaceable> > <replaceable>Array</replaceable> > <replaceable>Boolean</replaceable> > <replaceable>Number</replaceable> > <replaceable>String</replaceable> > <replaceable>Null</replaceable>

<replaceable>Object with n pairs</replaceable> > <replaceable>object with n - 1 pairs</replaceable>

<replaceable>Array with n elements</replaceable> > <replaceable>array with n - 1 elements</replaceable>
</synopsis>
      Objects with equal numbers of pairs are compared in the order:
<synopsis>
<replaceable>key-1</replaceable>, <replaceable>value-1</replaceable>, <replaceable>key-2</replaceable> ...
</synopsis>
      Note that object keys are compared in their storage order;
      in particular, since shorter keys are stored before longer keys, this
      can lead to results that might be unintuitive, such as:
<programlisting>
{ "aa": 1, "c": 1} > {"b": 1, "d": 1}
</programlisting>
      Similarly, arrays with equal numbers of elements are compared in the
      order:
<synopsis>
<replaceable>element-1</replaceable>, <replaceable>element-2</replaceable> ...
</synopsis>
      Primitive JSON values are compared using the same
      comparison rules as for the underlying
      <productname>PostgreSQL</productname> data type.  Strings are
      compared using the default database collation.
  </para>
 </sect2>

 <sect2 id="jsonb-subscripting">
  <title><type>jsonb</type> Subscripting</title>
  <para>
   The <type>jsonb</type> data type supports array-style subscripting expressions
   to extract and modify elements. Nested values can be indicated by chaining
   subscripting expressions, following the same rules as the <literal>path</literal>
   argument in the <literal>jsonb_set</literal> function. If a <type>jsonb</type>
   value is an array, numeric subscripts start at zero, and negative integers count
   backwards from the last element of the array. Slice expressions are not supported.
   The result of a subscripting expression is always of the jsonb data type.
  </para>

  <para>
   <command>UPDATE</command> statements may use subscripting in the
   <literal>SET</literal> clause to modify <type>jsonb</type> values. Subscript
   paths must be traversable for all affected values insofar as they exist. For
   instance, the path <literal>val['a']['b']['c']</literal> can be traversed all
   the way to <literal>c</literal> if every <literal>val</literal>,
   <literal>val['a']</literal>, and <literal>val['a']['b']</literal> is an
   object. If any <literal>val['a']</literal> or <literal>val['a']['b']</literal>
   is not defined, it will be created as an empty object and filled as
   necessary. However, if any <literal>val</literal> itself or one of the
   intermediary values is defined as a non-object such as a string, number, or
   <literal>jsonb</literal> <literal>null</literal>, traversal cannot proceed so
   an error is raised and the transaction aborted.
  </para>

  <para>
   An example of subscripting syntax:

<programlisting>

-- Extract object value by key
SELECT ('{"a": 1}'::jsonb)['a'];

-- Extract nested object value by key path
SELECT ('{"a": {"b": {"c": 1}}}'::jsonb)['a']['b']['c'];

-- Extract array element by index
SELECT ('[1, "2", null]'::jsonb)[1];

-- Update object value by key. Note the quotes around '1': the assigned
-- value must be of the jsonb type as well
UPDATE table_name SET jsonb_field['key'] = '1';

-- This will raise an error if any record's jsonb_field['a']['b'] is something
-- other than an object. For example, the value {"a": 1} has a numeric value
-- of the key 'a'.
UPDATE table_name SET jsonb_field['a']['b']['c'] = '1';

-- Filter records using a WHERE clause with subscripting. Since the result of
-- subscripting is jsonb, the value we compare it against must also be jsonb.
-- The double quotes make "value" also a valid jsonb string.
SELECT * FROM table_name WHERE jsonb_field['key'] = '"value"';
</programlisting>

   <type>jsonb</type> assignment via subscripting handles a few edge cases
   differently from <literal>jsonb_set</literal>. When a source <type>jsonb</type>
   value is <literal>NULL</literal>, assignment via subscripting will proceed
   as if it was an empty JSON value of the type (object or array) implied by the
   subscript key:

<programlisting>
-- Where jsonb_field was NULL, it is now {"a": 1}
UPDATE table_name SET jsonb_field['a'] = '1';

-- Where jsonb_field was NULL, it is now [1]
UPDATE table_name SET jsonb_field[0] = '1';
</programlisting>

   If an index is specified for an array containing too few elements,
   <literal>NULL</literal> elements will be appended until the index is reachable
   and the value can be set.

<programlisting>
-- Where jsonb_field was [], it is now [null, null, 2];
-- where jsonb_field was [0], it is now [0, null, 2]
UPDATE table_name SET jsonb_field[2] = '2';
</programlisting>

   A <type>jsonb</type> value will accept assignments to nonexistent subscript
   paths as long as the last existing element to be traversed is an object or
   array, as implied by the corresponding subscript (the element indicated by
   the last subscript in the path is not traversed and may be anything). Nested
   array and object structures will be created, and in the former case
   <literal>null</literal>-padded, as specified by the subscript path until the
   assigned value can be placed.

<programlisting>
-- Where jsonb_field was {}, it is now {"a": [{"b": 1}]}
UPDATE table_name SET jsonb_field['a'][0]['b'] = '1';

-- Where jsonb_field was [], it is now [null, {"a": 1}]
UPDATE table_name SET jsonb_field[1]['a'] = '1';
</programlisting>

  </para>
 </sect2>

 <sect2 id="datatype-json-transforms">
  <title>Transforms</title>

  <para>
   Additional extensions are available that implement transforms for the
   <type>jsonb</type> type for different procedural languages.
  </para>

  <para>
   The extensions for PL/Perl are called <literal>jsonb_plperl</literal> and
   <literal>jsonb_plperlu</literal>.  If you use them, <type>jsonb</type>
   values are mapped to Perl arrays, hashes, and scalars, as appropriate.
  </para>

  <para>
   The extension for PL/Python is called <literal>jsonb_plpython3u</literal>.
   If you use it, <type>jsonb</type> values are mapped to Python
   dictionaries, lists, and scalars, as appropriate.
  </para>

  <para>
   Of these extensions, <literal>jsonb_plperl</literal> is
   considered <quote>trusted</quote>, that is, it can be installed by
   non-superusers who have <literal>CREATE</literal> privilege on the
   current database.  The rest require superuser privilege to install.
  </para>
 </sect2>

 <sect2 id="datatype-jsonpath">
  <title>jsonpath Type</title>

  <indexterm zone="datatype-jsonpath">
   <primary>jsonpath</primary>
  </indexterm>

  <para>
   The <type>jsonpath</type> type implements support for the SQL/JSON path language
   in <productname>PostgreSQL</productname> to efficiently query JSON data.
   It provides a binary representation of the parsed SQL/JSON path
   expression that specifies the items to be retrieved by the path
   engine from the JSON data for further processing with the
   SQL/JSON query functions.
  </para>

  <para>
   The semantics of SQL/JSON path predicates and operators generally follow SQL.
   At the same time, to provide a natural way of working with JSON data,
   SQL/JSON path syntax uses some JavaScript conventions:
  </para>

  <itemizedlist>
   <listitem>
    <para>
     Dot (<literal>.</literal>) is used for member access.
    </para>
   </listitem>
   <listitem>
    <para>
     Square brackets (<literal>[]</literal>) are used for array access.
    </para>
   </listitem>
   <listitem>
    <para>
     SQL/JSON arrays are 0-relative, unlike regular SQL arrays that start from 1.
    </para>
   </listitem>
  </itemizedlist>

  <para>
   Numeric literals in SQL/JSON path expressions follow JavaScript rules,
   which are different from both SQL and JSON in some minor details.  For
   example, SQL/JSON path allows <literal>.1</literal> and
   <literal>1.</literal>, which are invalid in JSON.  Non-decimal integer
   literals and underscore separators are supported, for example,
   <literal>1_000_000</literal>, <literal>0x1EEE_FFFF</literal>,
   <literal>0o273</literal>, <literal>0b100101</literal>.  In SQL/JSON path
   (and in JavaScript, but not in SQL proper), there must not be an underscore
   separator directly after the radix prefix.
  </para>

  <para>
   An SQL/JSON path expression is typically written in an SQL query as an
   SQL character string literal, so it must be enclosed in single quotes,
   and any single quotes desired within the value must be doubled
   (see <xref linkend="sql-syntax-strings"/>).
   Some forms of path expressions require string literals within them.
   These embedded string literals follow JavaScript/ECMAScript conventions:
   they must be surrounded by double quotes, and backslash escapes may be
   used within them to represent otherwise-hard-to-type characters.
   In particular, the way to write a double quote within an embedded string
   literal is <literal>\"</literal>, and to write a backslash itself, you
   must write <literal>\\</literal>.  Other special backslash sequences
   include those recognized in JavaScript strings:
   <literal>\b</literal>,
   <literal>\f</literal>,
   <literal>\n</literal>,
   <literal>\r</literal>,
   <literal>\t</literal>,
   <literal>\v</literal>
   for various ASCII control characters,
   <literal>\x<replaceable>NN</replaceable></literal> for a character code
   written with only two hex digits,
   <literal>\u<replaceable>NNNN</replaceable></literal> for a Unicode
   character identified by its 4-hex-digit code point, and
   <literal>\u{<replaceable>N...</replaceable>}</literal> for a Unicode
   character code point written with 1 to 6 hex digits.
  </para>

  <para>
   A path expression consists of a sequence of path elements,
   which can be any of the following:
   <itemizedlist>
    <listitem>
     <para>
      Path literals of JSON primitive types:
      Unicode text, numeric, true, false, or null.
     </para>
    </listitem>
    <listitem>
     <para>
      Path variables listed in <xref linkend="type-jsonpath-variables"/>.
     </para>
    </listitem>
    <listitem>
     <para>
      Accessor operators listed in <xref linkend="type-jsonpath-accessors"/>.
     </para>
    </listitem>
    <listitem>
     <para>
      <type>jsonpath</type> operators and methods listed
      in <xref linkend="functions-sqljson-path-operators"/>.
     </para>
    </listitem>
    <listitem>
     <para>
      Parentheses, which can be used to provide filter expressions
      or define the order of path evaluation.
     </para>
    </listitem>
   </itemizedlist>
  </para>

  <para>
   For details on using <type>jsonpath</type> expressions with SQL/JSON
   query functions, see <xref linkend="functions-sqljson-path"/>.
  </para>

  <table id="type-jsonpath-variables">
   <title><type>jsonpath</type> Variables</title>
   <tgroup cols="2">
    <colspec colname="col1" colwidth="1*"/>
    <colspec colname="col2" colwidth="2*"/>
    <thead>
     <row>
      <entry>Variable</entry>
      <entry>Description</entry>
     </row>
    </thead>
    <tbody>
     <row>
      <entry><literal>$</literal></entry>
      <entry>A variable representing the JSON value being queried
      (the <firstterm>context item</firstterm>).
      </entry>
     </row>
     <row>
      <entry><literal>$varname</literal></entry>
      <entry>
        A named variable. Its value can be set by the parameter
        <parameter>vars</parameter> of several JSON processing functions;
        see <xref linkend="functions-json-processing-table"/> for details.
        <!-- TODO: describe PASSING clause once implemented !-->
      </entry>
     </row>
     <row>
      <entry><literal>@</literal></entry>
      <entry>A variable representing the result of path evaluation
      in filter expressions.
      </entry>
     </row>
    </tbody>
   </tgroup>
  </table>

  <table id="type-jsonpath-accessors">
   <title><type>jsonpath</type> Accessors</title>
   <tgroup cols="2">
    <colspec colname="col1" colwidth="1*"/>
    <colspec colname="col2" colwidth="2*"/>
    <thead>
     <row>
      <entry>Accessor Operator</entry>
      <entry>Description</entry>
     </row>
    </thead>
    <tbody>
     <row>
      <entry>
       <para>
        <literal>.<replaceable>key</replaceable></literal>
       </para>
       <para>
        <literal>."$<replaceable>varname</replaceable>"</literal>
       </para>
      </entry>
      <entry>
       <para>
        Member accessor that returns an object member with
        the specified key. If the key name matches some named variable
        starting with <literal>$</literal> or does not meet the
        JavaScript rules for an identifier, it must be enclosed in
        double quotes to make it a string literal.
       </para>
      </entry>
     </row>
     <row>
      <entry>
       <para>
        <literal>.*</literal>
       </para>
      </entry>
      <entry>
       <para>
        Wildcard member accessor that returns the values of all
        members located at the top level of the current object.
       </para>
      </entry>
     </row>
     <row>
      <entry>
       <para>
        <literal>.**</literal>
       </para>
      </entry>
      <entry>
       <para>
        Recursive wildcard member accessor that processes all levels
        of the JSON hierarchy of the current object and returns all
        the member values, regardless of their nesting level. This
        is a <productname>PostgreSQL</productname> extension of
        the SQL/JSON standard.
       </para>
      </entry>
     </row>
     <row>
      <entry>
       <para>
        <literal>.**{<replaceable>level</replaceable>}</literal>
       </para>
       <para>
        <literal>.**{<replaceable>start_level</replaceable> to
        <replaceable>end_level</replaceable>}</literal>
       </para>
      </entry>
      <entry>
       <para>
        Like <literal>.**</literal>, but selects only the specified
        levels of the JSON hierarchy. Nesting levels are specified as integers.
        Level zero corresponds to the current object. To access the lowest
        nesting level, you can use the <literal>last</literal> keyword.
        This is a <productname>PostgreSQL</productname> extension of
        the SQL/JSON standard.
       </para>
      </entry>
     </row>
     <row>
      <entry>
       <para>
        <literal>[<replaceable>subscript</replaceable>, ...]</literal>
       </para>
      </entry>
      <entry>
       <para>
        Array element accessor.
        <literal><replaceable>subscript</replaceable></literal> can be
        given in two forms: <literal><replaceable>index</replaceable></literal>
        or <literal><replaceable>start_index</replaceable> to <replaceable>end_index</replaceable></literal>.
        The first form returns a single array element by its index. The second
        form returns an array slice by the range of indexes, including the
        elements that correspond to the provided
        <replaceable>start_index</replaceable> and <replaceable>end_index</replaceable>.
       </para>
       <para>
        The specified <replaceable>index</replaceable> can be an integer, as
        well as an expression returning a single numeric value, which is
        automatically cast to integer. Index zero corresponds to the first
        array element. You can also use the <literal>last</literal> keyword
        to denote the last array element, which is useful for handling arrays
        of unknown length.
       </para>
      </entry>
     </row>
     <row>
      <entry>
       <para>
        <literal>[*]</literal>
       </para>
      </entry>
      <entry>
       <para>
        Wildcard array element accessor that returns all array elements.
       </para>
      </entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </sect2>
</sect1>