Internet-Draft | JSONPath | February 2023 |
Gössner, et al. | Expires 10 August 2023 | [Page] |
JSONPath defines a string syntax for selecting and extracting JSON (RFC 8259) values from a JSON value.¶
This note is to be removed before publishing as an RFC.¶
Status information for this document may be found at https://datatracker.ietf.org/doc/draft-ietf-jsonpath-base/.¶
Discussion of this document takes place on the JSON Path Working Group mailing list (mailto:jsonpath@ietf.org), which is archived at https://mailarchive.ietf.org/arch/browse/jsonpath/. Subscribe at https://www.ietf.org/mailman/listinfo/jsonpath/.¶
Source for this draft and an issue tracker can be found at https://github.com/ietf-wg-jsonpath/draft-ietf-jsonpath-base.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 10 August 2023.¶
Copyright (c) 2023 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
JSON [RFC8259] is a popular representation format for structured data values. JSONPath defines a string syntax for selecting and extracting JSON values from a JSON value.¶
JSONPath is not intended as a replacement for, but as a more powerful companion to, JSON Pointer [RFC6901]. See Appendix B.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
The grammatical rules in this document are to be interpreted as ABNF,
as described in [RFC5234].
ABNF terminal values in this document define Unicode code points rather than
their UTF-8 encoding.
For example, the Unicode PLACE OF INTEREST SIGN (U+2318) would be defined
in ABNF as %x2318
.¶
The terminology of [RFC8259] applies except where clarified below. The terms "Primitive" and "Structured" are used to group different kinds of values as in Section 1 of [RFC8259]; JSON Objects and Arrays are structured, all other values are primitive. Definitions for "Object", "Array", "Number", and "String" remain unchanged. Importantly "object" and "array" in particular do not take on a generic meaning, such as they would in a general programming context.¶
Additional terms used in this document are defined below.¶
As per [RFC8259], a structure conforming to the generic data model of JSON, i.e., composed of components such as structured values, namely JSON objects and arrays, and primitive data, namely numbers and text strings as well as the special values null, true, and false. [RFC8259] focuses on the textual representation of JSON values and does not fully define the value abstraction assumed here.¶
A name/value pair in an object. (A member is not itself a value.)¶
The name (a string) in a name/value pair constituting a member. This is also used in [RFC8259], but that specification does not formally define it. It is included here for completeness.¶
A value in a JSON array.¶
An integer that identifies a specific element in an array.¶
Short name for a JSONPath expression.¶
Short name for the value a JSONPath expression is applied to.¶
the position of a value within the argument. This can be thought of as a sequence of names and indexes navigating to the value through the objects and arrays in the argument, with the empty sequence indicating the argument itself. A location can be represented as a Normalized Path (defined below).¶
The pair of a value along with its location within the argument.¶
The unique node whose value is the entire argument.¶
The expression $
which refers to the root node of the argument.¶
The expression @
which refers to the current node in the context
of the evaluation of a filter expression (described later).¶
If the node is an array, the nodes of its elements. If the node is an object, the nodes of its member values. If the node is neither an array nor an object, it has no children.¶
The children of the node, together with the children of its children, and so forth recursively. More formally, the descendants relation between nodes is the transitive closure of the children relation.¶
The number of ancestors of the node within the value. The root node of the value has depth zero, the children of the root node have depth one, their children have depth two, and so forth.¶
One of the constructs which select children ([]
)
or descendants (..[]
) of an input value.¶
A list of nodes. While a nodelist can be represented in JSON, e.g. as an array, this document does not require or assume any particular representation.¶
A simple form of JSONPath expression that identifies a node in a value by providing a query that results in exactly that node. Similar to, but syntactically different from, a JSON Pointer [RFC6901].¶
Any Unicode [UNICODE] code point except high-surrogate and low-surrogate code points. In other words, integers in either of the inclusive base 16 ranges 0 to D7FF and E000 to 10FFFF. JSON string values are sequences of Unicode scalar values.¶
A nodelist containing at most one node.¶
A JSONPath expression built from segments each of which, regardless of the input value, produces a Singular Nodelist.¶
A single item within a segment that takes the input value and produces a nodelist consisting of child nodes of the input value.¶
This document models the argument as a tree of JSON values, each with its own node. A node is either the root node or one of its descendants.¶
This document models the result of applying a query to the argument as a nodelist (a list of nodes).¶
Nodes are the selectable parts of the argument. The only parts of an object that can be selected by a query are the member values. Member names and members (name/value pairs) cannot be selected. Thus, member values have nodes, but members and member names do not. Similarly, member values are children of an object, but members and member names are not.¶
This section is informative.¶
This document picks up Stefan Gössner's popular JSONPath proposal dated 2007-02-21 [JSONPath-orig], builds on the experience from the widespread deployment of its implementations, and provides a normative specification for it.¶
Appendix A describes how JSONPath was inspired by XML's XPath [XPath].¶
JSONPath was intended as a light-weight companion to JSON
implementations in programming languages such as PHP and JavaScript,
so instead of defining its own expression language, like XPath did,
JSONPath delegated parts of a query to the underlying
runtime, e.g., JavaScript's eval()
function.
As JSONPath was implemented in more environments, JSONPath
expressions became decreasingly portable.
For example, regular expression processing was often delegated to a
convenient regular expression engine.¶
This document aims to remove such implementation-specific dependencies and serve as a common JSONPath specification that can be used across programming languages and environments. This means that backwards compatibility is not always achieved; a design principle of this document is to go with a "consensus" between implementations even if it is rough, as long as that does not jeopardize the objective of obtaining a usable, stable JSON query language.¶
The JSON value a JSONPath query is applied to is, by definition, a valid JSON value. A JSON value is often constructed by parsing a JSON text.¶
The parsing of a JSON text into a JSON value and what happens if a JSON text does not represent valid JSON are not defined by this document. Sections 4 and 8 of [RFC8259] identify specific situations that may conform to the grammar for JSON texts but are not interoperable uses of JSON, as they may cause unpredictable behavior. This document does not attempt to define predictable behavior for JSONPath queries in these situations.¶
Specifically, the "Semantics" subsections of Sections 2.5.1, 2.5.2, 2.5.5, and 2.7.2 describe behavior that becomes unpredictable when the JSON value for one of the objects under consideration was constructed out of JSON text that exhibits multiple members for a single object that share the same member name ("duplicate names", see Section 4 of [RFC8259]). Also, selecting a child by name (Section 2.5.1) and comparing strings (Section "Comparisons" in Section 2.5.5) assume these strings are sequences of Unicode scalar values, becoming unpredictable if they are not (Section 8.2 of [RFC8259]).¶
This section is informative.¶
A JSONPath expression is applied to a JSON value, known as the argument. The output is a nodelist.¶
A JSONPath expression consists of an identifier followed by a series of zero or more segments each of which contains one or more selectors.¶
The root node identifier $
refers to the root node of the argument,
i.e., to the argument as a whole.¶
The current node identifier @
refers to the current node in the context
of the evaluation of a filter expression (described later).¶
Segments select children ([]
) or descendants (..[]
) of an input value.¶
Segments can use bracket notation, for example:¶
$['store']['book'][0]['title']¶
or the more compact dot notation, for example:¶
$.store.book[0].title¶
A JSONPath expression may use a combination of bracket and dot notations.¶
This document treats the bracket notations as canonical and defines the shorthand dot notation in terms of bracket notation. Examples and descriptions use shorthands where convenient.¶
A name selector, e.g. 'name'
, selects a named child of an object.¶
An index selector, e.g. 3
, selects an indexed child of an array.¶
A wildcard *
(Section 2.5.2) in the expression [*]
selects all children of a
node and in the expression ..[*]
selects all descendants of a node.¶
An array slice start:end:step
(Section 2.5.4) selects a series of
elements from an array, giving a start position, an end position, and
an optional step value that moves the position from the start to the
end.¶
Filter expressions ?<boolean expr>
select certain children of an object or array, as in:¶
$.store.book[?@.price < 10].title¶
Table 1 provides a brief overview of JSONPath syntax.¶
Syntax Element | Description |
---|---|
$
|
root node identifier (Section 2.4) |
@
|
current node identifier (Section 2.5.5) (valid only within filter selectors) |
[<selectors>]
|
child segment (Section 2.7.1) selects zero or more children of a node; contains one or more selectors, separated by commas |
.name
|
shorthand for ['name']
|
.*
|
shorthand for [*]
|
..[<selectors>]
|
descendant segment (Section 2.7.2): selects zero or more descendants of a node; contains one or more selectors, separated by commas |
..name
|
shorthand for ..['name']
|
..*
|
shorthand for ..[*]
|
'name'
|
name selector (Section 2.5.1): selects a named child of an object |
*
|
wildcard selector (Section 2.5.1): selects all children of a node |
3
|
index selector (Section 2.5.3): selects an indexed child of an array (from 0) |
0:100:5
|
array slice selector (Section 2.5.4): start:end:step for arrays |
?<expr>
|
filter selector (Section 2.5.5): selects particular children using a boolean expression |
length(@.foo)
|
function extension (Section 2.6): invokes a function in a filter expression |
This section is informative. It provides examples of JSONPath expressions.¶
The examples are based on the simple JSON value shown in Figure 1, representing a bookstore (that also has a bicycle).¶
{ "store": { "book": [ { "category": "reference", "author": "Nigel Rees", "title": "Sayings of the Century", "price": 8.95 }, { "category": "fiction", "author": "Evelyn Waugh", "title": "Sword of Honour", "price": 12.99 }, { "category": "fiction", "author": "Herman Melville", "title": "Moby Dick", "isbn": "0-553-21311-3", "price": 8.99 }, { "category": "fiction", "author": "J. R. R. Tolkien", "title": "The Lord of the Rings", "isbn": "0-395-19395-8", "price": 22.99 } ], "bicycle": { "color": "red", "price": 399 } } }
Table 2 shows some JSONPath queries that might be applied to this example and their intended results.¶
JSONPath | Intended result |
---|---|
$.store.book[*].author
|
the authors of all books in the store |
$..author
|
all authors |
$.store.*
|
all things in store, which are some books and a red bicycle |
$.store..price
|
the prices of everything in the store |
$..book[2]
|
the third book |
$..book[-1]
|
the last book in order |
$..book[0,1] $..book[:2]
|
the first two books |
$..book[?(@.isbn)]
|
all books with an ISBN number |
$..book[?(@.price<10)]
|
all books cheaper than 10 |
$..*
|
all member values and array elements contained in the input value |
A JSONPath expression is a string which, when applied to a JSON value, the argument, selects zero or more nodes of the argument and outputs these nodes as a nodelist.¶
A query MUST be encoded using UTF-8. The grammar for queries given in this document assumes that its UTF-8 form is first decoded into Unicode code points as described in [RFC3629]; implementation approaches that lead to an equivalent result are possible.¶
A string to be used as a JSONPath query needs to be well-formed and valid. A string is a well-formed JSONPath query if it conforms to the ABNF syntax in this document. A well-formed JSONPath query is valid if it also fulfills all semantic requirements posed by this document, which are:¶
A JSONPath implementation MUST raise an error for any query which is not well-formed and valid. The well-formedness and the validity of JSONPath queries are independent of the JSON value the query is applied to; no further errors relating to the well-formedness and the validity of a JSONPath query can be raised during application of the query to a value.¶
Obviously, an implementation can still fail when executing a JSONPath query, e.g., because of resource depletion, but this is not modeled in this document. However, the implementation MUST NOT silently malfunction. Specifically, if a valid JSONPath query is evaluated against a structured value whose size does not fit in the range of exact values, interfering with the correct interpretation of the query, the implementation MUST provide an indication of overflow.¶
(Readers familiar with the HTTP error model may be reminded of 400 type errors when pondering well-formedness and validity, while resource depletion and related errors are comparable to 500 type errors.)¶
Syntactically, a JSONPath query consists of a root identifier ($
), which
stands for a nodelist that contains the root node of the argument,
followed by a possibly empty sequence of segments.¶
json-path = root-identifier segments segments = *(S segment)¶
The syntax and semantics of segments are defined in Section 2.7.¶
In this document, the semantics of a JSONPath query define the required results and do not prescribe the internal workings of an implementation. This document may describe semantics in a procedural step-by-step fashion, but such descriptions are normative only in the sense that any implementation MUST produce an identical result, but not in the sense that implementors are required to use the same algorithms.¶
The semantics are that a valid query is executed against a value, the argument, and produces a nodelist (i.e., a list of zero or more nodes of the value).¶
The query is a root identifier followed by a sequence of zero or more segments, each of which is applied to the result of the previous root identifier or segment and provides input to the next segment. These results and inputs take the form of a nodelist.¶
Segments can be added to a query to drill further into the structure of the input value.¶
The nodelist resulting from the root identifier contains a single node, the argument. The nodelist resulting from the last segment is presented as the result of the query. Depending on the specific API, it might be presented as an array of the JSON values at the nodes, an array of Normalized Paths referencing the nodes, or both — or some other representation as desired by the implementation. Note that an empty nodelist is a valid query result.¶
A segment operates on each of the nodes in its input nodelist in turn, and the resultant nodelists are concatenated to produce the result of the segment. A node may be selected more than once and appears that number of times in the nodelist. Duplicate nodes are not removed.¶
A syntactically valid segment MUST NOT produce errors when executing the query. This means that some operations that might be considered erroneous, such as using an index lying outside the range of an array, simply result in fewer nodes being selected.¶
As a consequence of this approach, if any of the segments produces an empty nodelist, then the whole query produces an empty nodelist.¶
If a query may produce a nodelist with more than one possible ordering, a particular implementation may also produce distinct orderings in successive runs of the query.¶
Consider this example. With the argument {"a":[{"b":0},{"b":1},{"c":2}]}
, the
query $.a[*].b
selects the following list of nodes: 0
, 1
(denoted here by their value).¶
The query consists of $
followed by three segments: .a
, [*]
, and .b
.¶
Firstly, $
produces a nodelist consisting of just the argument.¶
Next, .a
selects from any object input node and selects the
node of any
member value of the input
node corresponding to the member name "a"
.
The result is again a list of one node: [{"b":0},{"b":1},{"c":2}]
.¶
Next, [*]
selects from any array input node all its elements
(for an object input node, it would select all its member
values, but not the member names).
The result is a list of three nodes: {"b":0}
, {"b":1}
, and {"c":2}
.¶
Finally, .b
selects from any object input node with a member name
b
and selects the node of the member value of the input node corresponding to that name.
The result is a list containing 0
, 1
.
This is the concatenation of three lists, two of length one containing
0
, 1
, respectively, and one of length zero.¶
Every JSONPath query (except those inside filter expressions, see Section 2.5.5) MUST begin with the root identifier $
.¶
root-identifier = "$"¶
Selectors appear only inside child segments (Section 2.7.1) and descendant segments (Section 2.7.2).¶
A selector produces a nodelist consisting of zero or more children of the input value.¶
There are various kinds of selectors which produce children of objects, children of arrays, or children of either objects or arrays.¶
selector = name-selector / wildcard-selector / slice-selector / index-selector / filter-selector¶
The syntax and semantics of each kind of selector are defined below.¶
A name selector '<name>'
selects at most one object member value.¶
In contrast to JSON, the JSONPath syntax allows strings to be enclosed in single or double quotes.¶
name-selector = string-literal string-literal = %x22 *double-quoted %x22 / ; "string" %x27 *single-quoted %x27 ; 'string' double-quoted = unescaped / %x27 / ; ' ESC %x22 / ; \" ESC escapable single-quoted = unescaped / %x22 / ; " ESC %x27 / ; \' ESC escapable ESC = %x5C ; \ backslash unescaped = %x20-21 / ; see RFC 8259 ; omit 0x22 " %x23-26 / ; omit 0x27 ' %x28-5B / ; omit 0x5C \ %x5D-10FFFF escapable = %x62 / ; b BS backspace U+0008 %x66 / ; f FF form feed U+000C %x6E / ; n LF line feed U+000A %x72 / ; r CR carriage return U+000D %x74 / ; t HT horizontal tab U+0009 "/" / ; / slash (solidus) U+002F "\" / ; \ backslash (reverse solidus) U+005C (%x75 hexchar) ; uXXXX U+XXXX hexchar = non-surrogate / (high-surrogate "\" %x75 low-surrogate) non-surrogate = ((DIGIT / "A"/"B"/"C" / "E"/"F") 3HEXDIG) / ("D" %x30-37 2HEXDIG ) high-surrogate = "D" ("8"/"9"/"A"/"B") 2HEXDIG low-surrogate = "D" ("C"/"D"/"E"/"F") 2HEXDIG HEXDIG = DIGIT / "A" / "B" / "C" / "D" / "E" / "F"¶
Note: double-quoted
strings follow the JSON string syntax (Section 7 of [RFC8259]);
single-quoted
strings follow an analogous pattern (Section "Syntax").
No attempt was made to improve on this syntax, so characters with
scalar values above 0x10000, such as U+1F914 ("🤔", THINKING FACE), need to be represented
by a pair of surrogate escapes ("\uD83E\uDD14"
in this case).¶
A name-selector
string MUST be converted to a
member name M by removing the surrounding quotes and
replacing each escape sequence with its equivalent Unicode character, as
in the table below:¶
Escape Sequence | Unicode Character | Description |
---|---|---|
\b | U+0008 | BS backspace |
\t | U+0009 | HT horizontal tab |
\n | U+000A | LF line feed |
\f | U+000C | FF form feed |
\r | U+000D | CR carriage return |
\" | U+0022 | quotation mark |
\' | U+0027 | apostrophe |
\/ | U+002F | slash (solidus) |
\\ | U+005C | backslash (reverse solidus) |
\uXXXX | U+XXXX | unicode character |
Applying the name-selector
to an object node
selects a member value whose name equals the member name M,
or selects nothing if there is no such member value.
Nothing is selected from a value that is not an object.¶
Note that processing the name selector requires comparing the member name string M with member name strings in the JSON to which the selector is being applied. Two strings MUST be considered equal if and only if they are identical sequences of Unicode scalar values. In other words, normalization operations MUST NOT be applied to either the member name string M from the JSONPath or to the member name strings in the JSON prior to comparison.¶
JSON:¶
{ "o": {"j j": {"k.k": 3}}, "'": {"@": 2} }¶
Queries:¶
The following examples show the name selector in use by child segments.¶
Query | Result | Result Paths | Comment |
---|---|---|---|
$.o['j j']['k.k']
|
3
|
$['o']['j j']['k.k']
|
Named value in nested object |
$.o["j j"]["k.k"]
|
3
|
$['o']['j j']['k.k']
|
Named value in nested object |
$["'"]["@"]
|
2
|
$['\'']['@']
|
Unusual member names |
A wildcard
selector selects the nodes of all children of an object or array.
The order in which the children of an object appear in the resultant nodelist is not stipulated,
since JSON objects are unordered.
Children of an array appear in array order in the resultant nodelist.¶
The wildcard
selector selects nothing from a primitive JSON value (that is,
a number, a string, true
, false
, or null
).¶
JSON:¶
{ "o": {"j": 1, "k": 2}, "a": [5, 3] }¶
Queries:¶
The following examples show the wildcard
selector in use by a child segment.¶
Query | Result | Result Paths | Comment |
---|---|---|---|
$[*]
|
{"j": 1, "k": 2} [5, 3]
|
$['o'] $['a']
|
Object values |
$.o[*]
|
1 2
|
$['o']['j'] $['o']['k']
|
Object values |
$.o[*]
|
2 1
|
$['o']['k'] $['o']['j']
|
Alternative result |
$.o[*, *]
|
1 2 2 1
|
$['o']['j'] $['o']['k'] $['o']['k'] $['o']['j']
|
Non-deterministic ordering |
$.a[*]
|
5 3
|
$['a'][0] $['a'][1]
|
Array members |
The example above with the query $.o[*, *]
shows that the wildcard selector may produce nodelists in distinct
orders each time it appears in the child segment, when it is applied to an object node with two or more
members (but not when it is applied to object nodes with fewer than two members or to array nodes).¶
An index selector <index>
matches at most one array element value.¶
index-selector = int ; decimal integer int = "0" / (["-"] DIGIT1 *DIGIT) ; - optional DIGIT1 = %x31-39 ; 1-9 non-zero digit¶
Applying the numerical index-selector
selects the corresponding
element. JSONPath allows it to be negative (see Section "Semantics").¶
To be valid, the index selector value MUST be in the I-JSON range of exact values, see Section 2.1.¶
Notes:
1. An index-selector
is an integer (in base 10, as in JSON numbers).
2. As in JSON numbers, the syntax does not allow octal-like integers with leading zeros such as 01
or -01
.¶
A non-negative index-selector
applied to an array selects an array element using a zero-based index.
For example, the selector 0
selects the first and the selector 4
selects the fifth element of a sufficiently long array.
Nothing is selected, and it is not an error, if the index lies outside the range of the array. Nothing is selected from a value that is not an array.¶
A negative index-selector
counts from the array end.
For example, the selector -1
selects the last and the selector -2
selects the penultimate element of an array with at least two elements.
As with non-negative indexes, it is not an error if such an element does
not exist; this simply means that no element is selected.¶
The array slice selector has the form <start>:<end>:<step>
.
It matches elements from arrays starting at index <start>
, ending at — but
not including — <end>
, while incrementing by step
with a default of 1
.¶
slice-selector = [start S] ":" S [end S] [":" [S step ]] start = int ; included in selection end = int ; not included in selection step = int ; default: 1 B = %x20 / ; Space %x09 / ; Horizontal tab %x0A / ; Line feed or New line %x0D ; Carriage return S = *B ; optional blank space¶
The slice selector consists of three optional decimal integers separated by colons.¶
To be valid, the integers provided MUST be in the I-JSON range of exact values, see Section 2.1.¶
The slice selector was inspired by the slice operator of ECMAScript 4 (ES4), which was deprecated in 2014, and that of Python.¶
This section is informative.¶
Array slicing is inspired by the behavior of the Array.prototype.slice
method
of the JavaScript language as defined by the ECMA-262 standard [ECMA-262],
with the addition of the step
parameter, which is inspired by the Python slice expression.¶
The array slice expression start:end:step
selects elements at indices starting at start
,
incrementing by step
, and ending with end
(which is itself excluded).
So, for example, the expression 1:3
(where step
defaults to 1
)
selects elements with indices 1
and 2
(in that order) whereas
1:5:2
selects elements with indices 1
and 3
.¶
When step
is negative, elements are selected in reverse order. Thus,
for example, 5:1:-2
selects elements with indices 5
and 3
, in
that order and ::-1
selects all the elements of an array in
reverse order.¶
When step
is 0
, no elements are selected.
(This is the one case that differs from the behavior of Python, which
raises an error in this case.)¶
The following section specifies the behavior fully, without depending on JavaScript or Python behavior.¶
A slice expression selects a subset of the elements of the input array, in
the same order
as the array or the reverse order, depending on the sign of the step
parameter.
It selects no nodes from a node that is not an array.¶
A slice is defined by the two slice parameters, start
and end
, and
an iteration delta, step
.
Each of these parameters is
optional. In the rest of this section, len
denotes the length of the input array.¶
The default value for step
is 1
.
The default values for start
and end
depend on the sign of step
,
as follows:¶
Condition | start | end |
---|---|---|
step >= 0 | 0 | len |
step < 0 | len - 1 | -len - 1 |
Slice expression parameters start
and end
are not directly usable
as slice bounds and must first be normalized.
Normalization for this purpose is defined as:¶
FUNCTION Normalize(i, len): IF i >= 0 THEN RETURN i ELSE RETURN len + i END IF¶
The result of the array index expression i
applied to an array
of length len
is the result of the array
slicing expression Normalize(i, len):Normalize(i, len)+1:1
.¶
Slice expression parameters start
and end
are used to derive slice bounds lower
and upper
.
The direction of the iteration, defined
by the sign of step
, determines which of the parameters is the lower bound and which
is the upper bound:¶
FUNCTION Bounds(start, end, step, len): n_start = Normalize(start, len) n_end = Normalize(end, len) IF step >= 0 THEN lower = MIN(MAX(n_start, 0), len) upper = MIN(MAX(n_end, 0), len) ELSE upper = MIN(MAX(n_start, -1), len-1) lower = MIN(MAX(n_end, -1), len-1) END IF RETURN (lower, upper)¶
The slice expression selects elements with indices between the lower and
upper bounds.
In the following pseudocode, a(i)
is the i+1
th element of the array a
(i.e., a(0)
is the first element, a(1)
the second, and so forth).¶
IF step > 0 THEN i = lower WHILE i < upper: SELECT a(i) i = i + step END WHILE ELSE if step < 0 THEN i = upper WHILE lower < i: SELECT a(i) i = i + step END WHILE END IF¶
When step = 0
, no elements are selected and the result array is empty.¶
JSON:¶
["a", "b", "c", "d", "e", "f", "g"]¶
Queries:¶
The following examples show the array slice selector in use by a child segment.¶
Query | Result | Result Paths | Comment |
---|---|---|---|
$[1:3]
|
"b" "c"
|
$[1] $[2]
|
Slice with default step |
$[1:5:2]
|
"b" "d"
|
$[1] $[3]
|
Slice with step 2 |
$[5:1:-2]
|
"f" "d"
|
$[5] $[3]
|
Slice with negative step |
$[::-1]
|
"g" "f" "e" "d" "c" "b" "a"
|
$[6] $[5] $[4] $[3] $[2] $[1] $[0]
|
Slice in reverse order |
The filter selector has the form ?<expr>
. It iterates over structured values, i.e., arrays and objects.¶
filter-selector = "?" S boolean-expr¶
During the iteration process the node of each array element or object member value being visited is known as the current node. A boolean expression, usually involving the current node, is evaluated and the current node is selected if and only if the expression yields true.¶
As the expression is composed of side-effect free components,
the order of evaluation does not need to be (and is not) defined.
Similarly, for conjunction (&&
) and disjunction (||
) (defined later), both a short-circuiting and a fully evaluating
implementation will lead to the same result; both implementation
strategies are therefore valid.¶
The current node is accessible via the current node identifier @
.
This identifier addresses the current node of the filter-selector that
is directly enclosing the identifier; note that within nested
filter-selectors, there is no syntax to address the current node of
any other than the directly enclosing filter-selector (i.e., of
filter-selectors enclosing the filter-selector that is directly
enclosing the identifier).¶
A test expression either tests the existence of a node
designated by an embedded query (see Section "Existence Tests") or tests the
result of a function expression (see Section 2.6).
In the latter case, if the function expression is of type
OptionalBoolean
or one of its subtypes, it tests whether the result
is true
; if the function expression is of type OptionalNodes
or
one of its subtypes, it tests whether the result is different from
Nothing
.¶
boolean-expr = logical-or-expr logical-or-expr = logical-and-expr *(S "||" S logical-and-expr) ; disjunction ; binds less tightly than conjunction logical-and-expr = basic-expr *(S "&&" S basic-expr) ; conjunction ; binds more tightly than disjunction basic-expr = paren-expr / comparison-expr / test-expr test-expr = [logical-not-op S] filter-path / ; path existence or non-existence function-expr ; OptionalBoolean or subtype filter-path = rel-path / json-path / function-expr ; OptionalNodes or subtype rel-path = current-node-identifier segments current-node-identifier = "@"¶
Parentheses MAY be used within boolean-expr
for grouping.¶
paren-expr = [logical-not-op S] "(" S boolean-expr S ")" ; parenthesized expression logical-not-op = "!" ; logical NOT operator¶
Comparisons are between primitive values (that is, numbers, strings, true
, false
,
and null
),
Singular Paths, each of which selects at most one node, and
function expressions (see Section 2.6) that
return a primitive value or at most one node.¶
comparison-expr = comparable S comparison-op S comparable comparable = number / string-literal / true / false / null / singular-path / ; Singular Path value function-expr ; OptionalNodeOrValue or subtype comparison-op = "==" / "!=" / "<=" / ">=" / "<" / ">" singular-path = rel-singular-path / abs-singular-path rel-singular-path = current-node-identifier singular-path-segments abs-singular-path = root-identifier singular-path-segments singular-path-segments = *(S (name-segment / index-segment)) name-segment = ("[" name-selector "]") / ("." member-name-shorthand) index-segment = "[" index-selector "]"¶
Alphabetic characters in ABNF are case-insensitive, so "e" can be either "e" or "E".¶
true
, false
, and null
are lower-case only (case-sensitive).¶
number = (int / "-0") [ frac ] [ exp ] ; decimal number frac = "." 1*DIGIT ; decimal fraction exp = "e" [ "-" / "+" ] 1*DIGIT ; decimal exponent true = %x74.72.75.65 ; true false = %x66.61.6c.73.65 ; false null = %x6e.75.6c.6c ; null¶
The following table lists filter expression operators in order of precedence from highest (binds most tightly) to lowest (binds least tightly).¶
Precedence | Operator type | Syntax |
---|---|---|
5 | Grouping |
(...)
|
4 | Logical NOT |
!
|
3 | Relations |
== != < <= > >=
|
2 | Logical AND |
&&
|
1 | Logical OR |
||
|
The filter selector works with arrays and objects exclusively. Its result is a list of zero, one, multiple or all of their array elements or member values, respectively. Applied to primitive values, it selects nothing.¶
The order in which the children of an object appear in the resultant nodelist is not stipulated, since JSON objects are unordered. Children of an array appear in array order in the resultant nodelist.¶
A path by itself in a Boolean context is an existence test which yields true if the path selects at least one node and yields false if the path does not select any nodes.¶
Existence tests differ from comparisons in that:¶
To test the value of a node selected by a path, an explicit comparison is necessary.
For example, to test whether the node selected by the path @.foo
has the value null
, use @.foo == null
(see Section 2.8)
rather than the negated existence test !@.foo
(which yields false if @.foo
selects a node, regardless of the node's value).¶
The comparison operators ==
and <
are defined first and then these are used to define !=
, <=
, >
, and >=
.¶
When a path resulting in an empty nodelist appears on either side of a comparison:¶
==
yields true if and only if the comparison
is between two paths each of which result in an empty nodelist.¶
<
yields false.¶
When any path on either side of a comparison results in a nodelist consisting of a single node, each such path is replaced by the value of its node and then:¶
a comparison using the operator ==
yields true if and only if the comparison
is between:¶
equal objects with no duplicate names, that is where:¶
a comparison using the operator <
yields true if and only if
the comparison is between values which are both numbers or both strings and which satisfy the comparison:¶
Note that comparisons using the operator <
yield false if either value being
compared is an object, array, boolean, or null
.¶
!=
, <=
, >
, and >=
are defined in terms of the other comparison operators. For any a
and b
:¶
a != b
yields true if and only if a == b
yields false.¶
a <= b
yields true if and only if a < b
yields true or a == b
yields true.¶
a > b
yields true if and only if b < a
yields true.¶
a >= b
yields true if and only if b < a
yields true or a == b
yields true.¶
The logical AND, OR, and NOT operators have the normal semantics of Boolean algebra and obey its laws (see, for example, [BOOLEAN-LAWS]).¶
Filter selectors may use function extensions, which are covered in Section 2.6.¶
The first set of examples shows some comparison expressions and their result with a given JSON value as input.¶
JSON:¶
{ "obj": {"x": "y"}, "arr": [2, 3] }¶
Comparisons:¶
Comparison | Result | Comment |
---|---|---|
$.absent1 == $.absent2
|
true | Empty nodelists |
$.absent1 <= $.absent2
|
true |
== implies <=
|
$.absent == 'g'
|
false | Empty nodelist |
$.absent1 != $.absent2
|
false | Empty nodelists |
$.absent != 'g'
|
true | Empty nodelist |
1 <= 2
|
true | Numeric comparison |
1 > 2
|
false | Strict, numeric comparison |
13 == '13'
|
false | Type mismatch |
'a' <= 'b'
|
true | String comparison |
'a' > 'b'
|
false | Strict, string comparison |
$.obj == $.arr
|
false | Type mismatch |
$.obj != $.arr
|
true | Type mismatch |
$.obj == $.obj
|
true | Object comparison |
$.obj != $.obj
|
false | Object comparison |
$.arr == $.arr
|
true | Array comparison |
$.arr != $.arr
|
false | Array comparison |
$.obj == 17
|
false | Type mismatch |
$.obj != 17
|
true | Type mismatch |
$.obj <= $.arr
|
false | Objects and arrays are not ordered |
$.obj < $.arr
|
false | Objects and arrays are not ordered |
$.obj <= $.obj
|
true |
== implies <=
|
$.arr <= $.arr
|
true |
== implies <=
|
1 <= $.arr
|
false | Arrays are not ordered |
1 >= $.arr
|
false | Arrays are not ordered |
1 > $.arr
|
false | Arrays are not ordered |
1 < $.arr
|
false | Arrays are not ordered |
true <= true
|
true |
== implies <=
|
true > true
|
false | Booleans are not ordered |
The second set of examples shows some complete JSONPath queries that make use of filter selectors, and the results of evaluating these queries on a given JSON value as input. (Note that two of the queries employ function extensions; please see Sections 2.6.5 and 2.6.6 below for details about these.)¶
JSON:¶
{ "a": [3, 5, 1, 2, 4, 6, {"b": "j"}, {"b": "k"}, {"b": {}}, {"b": "kilo"}], "o": {"p": 1, "q": 2, "r": 3, "s": 5, "t": {"u": 6}}, "e": "f" }¶
Queries:¶
The following examples show the filter selector in use by a child segment.¶
Query | Result | Result Paths | Comment |
---|---|---|---|
$.a[?@.b == 'kilo']
|
{"b": "kilo"}
|
$['a'][9]
|
Member value comparison |
$.a[?@>3.5]
|
5 4 6
|
$['a'][1] $['a'][4] $['a'][5]
|
Array value comparison |
$.a[?@.b]
|
{"b": "j"} {"b": "k"} {"b": {}} {"b": "kilo"}
|
$['a'][6] $['a'][7] $['a'][8] $['a'][9]
|
Array value existence |
$[?@.*]
|
[3, 5, 1, 2, 4, 6, {"b": "j"}, {"b": "k"}, {"b": {}}, {"b": "kilo"}] {"p": 1, "q": 2, "r": 3, "s": 5, "t": {"u": 6}}
|
$['a'] $['o']
|
Existence of non-singular paths |
$[?@[?@.b]]
|
[3, 5, 1, 2, 4, 6, {"b": "j"}, {"b": "k"}, {"b": {}}, {"b": "kilo"}]
|
$['a']
|
Nested filters |
$.o[?@<3, ?@<3]
|
1 2 2 1
|
$['o']['p'] $['o']['q'] $['o']['q'] $['o']['p']
|
Non-deterministic ordering |
$.a[?@<2 || @.b == "k"]
|
1 {"b": "k"}
|
$['a'][2] $['a'][7]
|
Array value logical OR |
$.a[?match(@.b, "[jk]")]
|
{"b": "j"} {"b": "k"}
|
$['a'][6] $['a'][7]
|
Array value regular expression match |
$.a[?search(@.b, "[jk]")]
|
{"b": "j"} {"b": "k"} {"b": "kilo"}
|
$['a'][6] $['a'][7] $['a'][9]
|
Array value regular expression search |
$.o[?@>1 && @<4]
|
2 3
|
$['o']['q'] $['o']['r']
|
Object value logical AND |
$.o[?@>1 && @<4]
|
3 2
|
$['o']['r'] $['o']['q']
|
Alternative result |
$.o[?@.u || @.x]
|
{"u": 6}
|
$['o']['t']
|
Object value logical OR |
$.a[?(@.b == $.x)]
|
3 5 1 2 4 6
|
$['a'][0] $['a'][1] $['a'][2] $['a'][3] $['a'][4]
|
Comparison of paths with no values |
$.a[?(@ == @)]
|
3 5 1 2 4 6 {"b": "j"} {"b": "k"} {"b": {}} {"b": "kilo"}
|
$['a'][0] $['a'][1] $['a'][2] $['a'][3] $['a'][4] $['a'][5] $['a'][6] $['a'][7] $['a'][8] $['a'][9]
|
Comparisons of primitive and of structured values |
The example above with the query $.o[?@<3, ?@<3]
shows that a filter selector may produce nodelists in distinct
orders each time it appears in the child segment.¶
Beyond the filter expression functionality defined in the preceding subsections, JSONPath defines an extension point that can be used to add filter expression functionality: "Function Extensions".¶
This section defines the extension point as well as four function extensions that use this extension point. While these mechanisms are designed to use the extension point, they are an integral part of the JSONPath specification and are mandatory to implement.¶
A function extension defines a registered name (see Section 3.2) that can be applied to a sequence of zero or more arguments, producing a result.¶
A function extension MUST be defined such that its evaluation is side-effect free, i.e., all possible orders of evaluation and choices of short-circuiting or full evaluation of an expression containing it must lead to the same result. (Note that memoization or logging are not side effects in this sense as they are visible at the implementation level only — they do not influence the result of the evaluation.)¶
function-name = function-name-first *function-name-char function-name-first = LCALPHA function-name-char = function-name-first / "_" / DIGIT LCALPHA = %x61-7A ; "a".."z" function-expr = function-name "(" S [function-argument *(S "," S function-argument)] S ")" function-argument = filter-path / comparable¶
A function argument is a filter-path
or a comparable
.¶
According to Section 2.5.5, a function-expr
is valid as a filter-path
or a comparable
.¶
Any function expressions in a query must be well-formed (by conforming to the above ABNF) and correctly typed, otherwise the JSONPath implementation MUST raise an error (see Section 2.1). To define which function expressions are correctly typed, a type system is first introduced.¶
Each argument and result of a function extension must have a declared type.¶
A type is a set of instances. A type is a subtype of another type if its set of instances (possibly after coercion) is a subset of the set of instances of the other type.¶
Table 13 defines the available types in terms of abstract instances, where n
denotes a node, v
denotes a value, and nl
denotes
a non-empty nodelist. The table also lists the subtypes of each type.¶
Type | Abstract Instances | Subtypes |
---|---|---|
OptionalNodeOrValue
|
Node(n) , Value(v) , Nothing
|
OptionalNode , OptionalValue , Value
|
OptionalNode
|
Node(n) , Nothing
|
|
OptionalValue
|
Value(v) , Nothing
|
Value , OptionalBoolean
|
Value
|
Value(v)
|
Boolean
|
OptionalBoolean
|
Value(true) , Value(false) , Nothing
|
Boolean
|
Boolean
|
Value(true) , Value(false)
|
|
OptionalNodes
|
Nodes(nl) , Nothing
|
OptionalNode
|
Notes:¶
OptionalNodeOrValue
is an abstraction of a comparable
(which may appear on either side of a comparison or as a function argument).¶
OptionalNode
is an abstraction of a Singular Path.¶
Value
is an abstraction of a primitive value.¶
Boolean
is an abstraction of a primitive value that is either
true
or false
.¶
OptionalValue
is an abstraction of a primitive value that may
alternatively be absent (Nothing
).¶
OptionalNodes
is an abstraction of a filter-path
(which appears
in a test expression or as a function argument).¶
The abstract instances above can be obtained from the concrete representations in Table 14.¶
Abstract Instance | Concrete Representations |
---|---|
Node(n)
|
Singular Path resulting in a nodelist containing just the node n
|
Value(v)
|
JSON value v
|
Nothing
|
Singular Path or filter-path resulting in an empty nodelist |
Nodes(nl)
|
filter-path resulting in the non-empty nodelist nl
|
The following subtype relationships depend on coercion:¶
OptionalNode
is a subtype of OptionalValue
via coercion since the OptionalNode
instance Node(n)
can be coerced to
the OptionalValue
instance Value(v)
, where v
is the value of the node n
.¶
OptionalNode
is a subtype of OptionalNodes
via coercion since the OptionalNode
instance Node(n)
can be coerced to
the OptionalNodes
instance Nodes(l)
, where l
is a nodelist consisting of just the node n
.¶
The type correctness of function expressions can now be defined in terms of this type system.¶
A function expression is correctly typed if all the following are true:¶
filter-path
in a test expression, the function
is defined to have result type OptionalNodes
or one of its subtypes,
or to have result type OptionalBoolean
or one of its subtypes.¶
comparable
in a comparison, the function
is defined to have result type OptionalNodeOrValue
or one of its subtypes.¶
For it and any function expression it contains, each argument of the function matches the defined type of the argument according to one of the following rules:¶
Value
or any type of which Value
is a subtype.¶
OptionalNode
or any type of which OptionalNode
is a subtype.¶
filter-path
or a Singular Path and the defined type of the argument is OptionalNodes
.¶
length
Function Extension
The "length" function extension provides a way to compute the length of a value and make that available for further processing in the filter expression:¶
$[?length(@.authors) >= 5]¶
Its only argument is an optional value (possibly taken from a singular path as in the example above). The result also is an optional value, an unsigned integer.¶
Nothing
.¶
count
Function Extension
The "count" function extension provides a way to obtain the number of nodes in a nodelist and make that available for further processing in the filter expression:¶
$[?count(@.*.author) >= 5]¶
Its only argument is a nodelist. The result is a value, an unsigned integer, that gives the number of nodes in the nodelist. Note that there is no deduplication of the nodelist.¶
match
Function Extension
OptionalNodeOrValue
(string)¶
Value
(string conforming to [I-D.draft-ietf-jsonpath-iregexp])¶
OptionalBoolean
(true
, false
, or Nothing
)¶
The "match" function extension provides a way to check whether (the entirety of, see Section 2.6.6 below) a given string matches a given regular expression, which is in [I-D.draft-ietf-jsonpath-iregexp] form.¶
$[?match(@.date, "1974-05-..")]¶
Its first argument is an optional string that is matched against the iregexp
contained in the string that is the second argument.
The result is true
if the string matches the iregexp and false
otherwise.¶
The result is Nothing
if the first argument is not a string or
the second argument is not a string conforming to [I-D.draft-ietf-jsonpath-iregexp].¶
search
Function Extension
OptionalNodeOrValue
(string)¶
Value
(string conforming to [I-D.draft-ietf-jsonpath-iregexp])¶
OptionalBoolean
(true
, false
, or Nothing
)¶
The "search" function extension provides a way to check whether a given string contains a substring that matches a given regular expression, which is in [I-D.draft-ietf-jsonpath-iregexp] form.¶
$[?search(@.author, "[BR]ob")]¶
Its first argument is an optional string that is searched for at least one
substring that matches the iregexp contained in the string
that is the second argument.
The result is true
if such a substring exists, false
otherwise.¶
The result is Nothing
if the first argument is not a string or
the second argument is not a string conforming to [I-D.draft-ietf-jsonpath-iregexp].¶
Query | Comment |
---|---|
$[?length(@) < 3]
|
Valid typing |
$[?length(@.*) < 3]
|
Invalid typing since @.* is a non-singular path |
$[?count(@.*) == 1]
|
Valid typing |
$[?count(1) == 1]
|
Invalid typing since 1 is not a path |
$[?count(foo(@.*)) == 1]
|
Valid typing, where foo is a function extension with argument of type OptionalNodes and result type OptionalNodes
|
$[?match(@.timezone, 'Europe/.*')]
|
Valid typing |
$[?match(@.timezone, 'Europe/.*') == true]
|
Invalid typing since OptionBoolean is not a comparable
|
Segments apply one or more selectors to an input value and concatenate the results into a single nodelist.¶
It turns out that the more segments there are in a query, the greater the depth in the input value of the nodes of the resultant nodelist:¶
There are two kinds of segment: child segments and descendant segments.¶
segment = child-segment / descendant-segment¶
The syntax and semantics of each kind of segment are defined below.¶
The child segment consists of a non-empty, comma-separated sequence of selectors enclosed in square brackets.¶
Shorthand notations are also provided for when there is a single wildcard or name selector.¶
child-segment = bracketed-selection / ("." (wildcard-selector / member-name-shorthand)) bracketed-selection = "[" S selector *(S "," S selector) S "]" member-name-shorthand = name-first *name-char name-first = ALPHA / "_" / %x80-10FFFF ; any non-ASCII Unicode character name-char = DIGIT / name-first DIGIT = %x30-39 ; 0-9 ALPHA = %x41-5A / %x61-7A ; A-Z / a-z¶
.*
, a child-segment
directly built from a wildcard-selector
, is
shorthand for [*]
.¶
.<member-name>
, a child-segment
built from a
member-name-shorthand
, is shorthand for ['<member-name>']
.
Note that this can only be used with member names that are composed of certain
characters, as specified in the ABNF rule member-name-shorthand
.
Thus, for example, $.foo.bar
is shorthand for $['foo']['bar']
(but not for $['foo.bar']
).¶
A child segment contains a sequence of selectors, each of which selects zero or more children of the input value.¶
Selectors of different kinds may be combined within a single child segment.¶
The resulting nodelist of a child segment is the concatenation of the nodelists from each of its selectors in the order that the selectors appear in the list. Note that any node matched by more than one selector is kept as many times in the nodelist.¶
Where a selector can produce a nodelist in more than one possible order, each occurrence of the selector in the child segment may evaluate to produce a nodelist in a distinct order.¶
So a child segment drills down one more level into the structure of the input value.¶
The descendant segment consists of a double dot ..
followed by a child segment (using bracket notation).¶
Shortand notations are also provided that correspond to the shorthand forms of the child segment.¶
descendant-segment = ".." (bracketed-selection / wildcard-selector / member-name-shorthand)¶
..*
, the descendant-segment
directly built from a
wildcard-selector
, is shorthand for ..[*]
.¶
..<member-name>
, a descendant-segment
built from a
member-name-shorthand
, is shorthand for ..['<member-name>']
.
As with the similar shorthand of a child-segment
, note that this can
only be used with member names that are composed of certain
characters, as specified in the ABNF rule member-name-shorthand
.¶
Note that ..
on its own is not a valid segment.¶
A descendant segment produces zero or more descendants of the input value.¶
A descendant selector visits the input value and each of its descendants such that:¶
The order in which the children of an object are visited is not stipulated, since JSON objects are unordered.¶
Suppose the descendant segment is of the form ..[<selectors>]
(after converting any shorthand
form to bracket notation)
and the nodes, in the order visited, are D1
, ..., Dn
(where n >= 1
).
Note that D1
is the input value.¶
For each i
such that 1 <= i <= n
, the nodelist Ri
is defined to be a result of applying
the child segment [<selectors>]
to the node Di
.¶
The result of the descendant selector is the concatenation of R1
, ..., Rn
(in that order).¶
So a descendant segment drills down one or more levels into the structure of the input value.¶
JSON:¶
{ "o": {"j": 1, "k": 2}, "a": [5, 3, [{"j": 4}, {"k": 6}]] }¶
Queries:¶
Query | Result | Result Paths | Comment |
---|---|---|---|
$..j
|
1 4
|
$['o']['j'] $['a'][2][0]['j']
|
Object values |
$..j
|
4 1
|
$['a'][2][0]['j'] $['o']['j']
|
Alternative result |
$..[0]
|
5 {"j": 4}
|
$['a'][0] $['a'][2][0]
|
Array values |
$..[0]
|
{"j": 4} 5
|
$['a'][2][0] $['a'][0]
|
Alternative result |
$..[*] $..*
|
{"j": 1, "k" : 2} [5, 3, [{"j": 4}, {"k": 6}]] 1 2 5 3 [{"j": 4}, {"k": 6}] {"j": 4} 4 {"k": 6} 6
|
$['o'] $['a'] $['o']['j'] $['o']['k'] $['a'][0] $['a'][1] $['a'][2] $['a'][2][0] $['a'][2][0]['j'] $['a'][2][1] $['a'][2][1]['k']
|
All values |
$..o
|
{"j": 1, "k": 2}
|
$['o']
|
Input value is visited |
$.o..[*, *]
|
1 2 2 1
|
$['o']['j'] $['o']['k'] $['o']['k'] $['o']['j']
|
Non-deterministic ordering |
$.a..[0, 1]
|
5 3 {"j": 4} {"k": 6}
|
$['a'][0] $['a'][1] $['a'][0][2][0] $['a'][0][2][1]
|
Multiple segments |
Note: The ordering of the results for the $..[*]
and $..*
examples above is not guaranteed, except that:¶
{"j": 1, "k": 2}
must appear before 1
and 2
,¶
[5, 3, [{"j": 4}, {"k": 6}]]
must appear before 5
, 3
, and [{"j": 4}, {"k": 6}]
,¶
5
must appear before 3
which must appear before [{"j": 4}, {"k": 6}]
,¶
5
and 3
must appear before {"j": 4}
, 4
, , {"k": 6}
, and 6
,¶
[{"j": 4}, {"k": 6}]
must appear before {"j": 4}
and {"k": 6}
,¶
{"j": 4}
must appear before 4
, and¶
{"k": 6}
must appear before 6
.¶
The example above with the query $.o..[*, *]
shows that a selector may produce nodelists in distinct orders
each time it appears in the descendant segment.¶
The example above with the query $.a..[0, 1]
shows that the child segment [0, 1]
is applied to each node
in turn (rather than the nodes being visited once per selector, which is the case for some JSONPath implementations
that do not conform to this specification).¶
null
Note that JSON null
is treated the same as any other JSON value: it is not taken to mean "undefined" or "missing".¶
JSON:¶
{"a": null, "b": [null], "c": [{}], "null": 1}¶
Queries:¶
Query | Result | Result Paths | Comment |
---|---|---|---|
$.a
|
null
|
$['a']
|
Object value |
$.a[0]
|
null used as array |
||
$.a.d
|
null used as object |
||
$.b[0]
|
null
|
$['b'][0]
|
Array value |
$.b[*]
|
null
|
$['b'][0]
|
Array value |
$.b[?@]
|
null
|
$['b'][0]
|
Existence |
$.b[?@==null]
|
null
|
$['b'][0]
|
Comparison |
$.c[?(@.d==null)]
|
Comparison with "missing" value | ||
$.null
|
1
|
$['null']
|
Not JSON null at all, just a member name string |
A Normalized Path is a canonical representation of the location of a node in a value and
uniquely identifies the node in the value.
Specifically, a Normalized Path is a JSONPath query with restricted syntax (defined below),
e.g., $['book'][3]
, which when applied to the value results in a nodelist consisting
of just the node identified by the Normalized Path.
Note that a Normalized Path represents the identity of a node in a specific value.
There is precisely one Normalized Path identifying any particular node in a value.¶
A canonical representation of a nodelist is as a JSON arrays of strings, where the strings are Normalized Paths.¶
Normalized Paths provide a predictable format that simplifies testing and post-processing of nodelists, e.g., to remove duplicate nodes. Normalized Paths are used in this document as result paths in examples.¶
Normalized Paths use the canonical bracket notation, rather than dot notation.¶
Single quotes are used to delimit string member names. This reduces the number of characters that need escaping when Normalized Paths appear in double quote delimited strings, e.g., in JSON texts.¶
Certain characters are escaped, in one and only one way; all other characters are unescaped.¶
Note: Normalized Paths are Singular Paths, but not all Singular Paths are Normalized Paths.
For example, $[-3]
is a Singular Path, but is not a Normalized Path.
The Normalized Path equivalent to $[-3]
would have an index equal to the array length minus 3
.
(The array length must be at least 3
if $[-3]
is to identify a node.)¶
normalized-path = root-identifier *(normal-index-segment) normal-index-segment = "[" normal-selector "]" normal-selector = normal-name-selector / normal-index-selector normal-name-selector = %x27 *normal-single-quoted %x27 ; 'string' normal-single-quoted = normal-unescaped / ESC normal-escapable normal-unescaped = ; omit %x0-1F control codes %x20-26 / ; omit 0x27 ' %x28-5B / ; omit 0x5C \ %x5D-10FFFF normal-escapable = %x62 / ; b BS backspace U+0008 %x66 / ; f FF form feed U+000C %x6E / ; n LF line feed U+000A %x72 / ; r CR carriage return U+000D %x74 / ; t HT horizontal tab U+0009 "'" / ; ' apostrophe U+0027 "\" / ; \ backslash (reverse solidus) U+005C (%x75 normal-hexchar) ; certain values u00xx U+00XX normal-hexchar = "0" "0" ( ("0" %x30-37) / ; "00"-"07" ; omit U+0008-U+000A BS HT LF ("0" %x62) / ; "0b" ; omit U+000C-U+000D FF CR ("0" %x65-66) / ; "0e"-"0f" ("1" normal-HEXDIG) ) normal-HEXDIG = DIGIT / %x61-66 ; "0"-"9", "a"-"f" normal-index-selector = "0" / (DIGIT1 *DIGIT) ; non-negative decimal integer¶
Since there can only be one Normalized Path identifying a given node, the syntax
stipulates which characters are escaped and which are not.
So the definition of normal-hexchar
is designed for hex escaping of characters
which are not straightforwardly-printable, for example U+000B LINE TABULATION, but
for which no standard JSON escape, such as \n
, is available.¶
Path | Normalized Path | Comment |
---|---|---|
$.a
|
$['a']
|
Object value |
$[1]
|
$[1]
|
Array index |
$[-3]
|
$[2]
|
Negative array index for an array of length 5 |
$.a.b[1:2]
|
$['a']['b'][1]
|
Nested structure |
$["\u000B"]
|
$['\u000b']
|
Unicode escape |
$["\u0061"]
|
$['a']
|
Unicode character |
IANA is requested to register the following media type [RFC6838]:¶
application¶
jsonpath¶
N/A¶
N/A¶
binary (UTF-8)¶
See the Security Considerations section of RFCXXXX.¶
N/A¶
RFCXXXX¶
Applications that need to convey queries in JSON data¶
N/A¶
Person & email address to contact for further information: iesg@ietf.org¶
This specification defines a new "Function Extensions sub-registry" in a new "JSONPath Parameters registry", with the policy "expert review" (Section 4.5 of [BCP26]).¶
The experts are instructed to be frugal in the allocation of function extension names that are suggestive of generally applicable semantics, keeping them in reserve for functions that are likely to enjoy wide use and can make good use of their conciseness. The expert is also instructed to direct the registrant to provide a specification (Section 4.6 of [BCP26]), but can make exceptions, for instance when a specification is not available at the time of registration but is likely forthcoming. If the expert becomes aware of function extensions that are deployed and in use, they may also initiate a registration on their own if they deem such a registration can avert potential future collisions.¶
Each entry in the registry must include:¶
a lower case ASCII [STD80] string that starts with a letter and can
contain letters, digits and underscore characters afterwards
([a-z][_a-z0-9]*
).¶
a brief description¶
A comma-separated list of zero or more types of the arguments expected for this function extension¶
The type of the result for this function extension¶
(see Section 2.3 of [BCP26])¶
a reference document that provides a description of the function extension¶
Initial entries in this sub-registry are as listed in Table 20; the Column "Change Controller" always has the value "IESG" and the column "Reference" always has the value "Section 2.6 of RFCthis":¶
Function Name | Brief description | Input | Output |
---|---|---|---|
length | length of array |
OptionalValue
|
OptionalValue
|
count | size of nodelist |
OptionalNodes
|
Value
|
match | regular expression full match |
OptionalNodeOrValue , Value
|
OptionalBoolean
|
search | regular expression substring match |
OptionalNodeOrValue , Value
|
OptionalBoolean
|
Security considerations for JSONPath can stem from¶
Historically, JSONPath has often been implemented by feeding parts of
the query to an underlying programming language engine, e.g.,
JavaScript's eval()
function.
This approach is well known to lead to injection attacks and would
require perfect input validation to prevent these attacks (see
Section 12 of [RFC8259] for similar considerations for JSON itself).
Instead, JSONPath implementations need to implement the entire syntax
of the query without relying on the parsers of programming language
engines.¶
Attacks on availability may attempt to trigger unusually expensive runtime performance exhibited by certain implementations in certain cases. (See Section 10 of [RFC8949] for issues in hash-table implementations, and Section 8 of [I-D.draft-ietf-jsonpath-iregexp] for performance issues in regular expression implementations.) Implementers need to be aware that good average performance is not sufficient as long as an attacker can choose to submit specially crafted JSONPath queries or arguments that trigger surprisingly high, possibly exponential, CPU usage or, for example via a naive recursive implementation of the descendant segment, stack overflow. Implementations need to have appropriate resource management to mitigate these attacks.¶
JSONPath queries are often not static, but formed from variables that
provide index values, member names, or values to compare with in a
filter expression.
These variables need to be translated into the form they take in a
JSONPath query, e.g., by escaping string delimiters, or by only
allowing specific constructs such as .name
to be formed when the
given values allow that.
Failure to perform these translations correctly can lead to unexpected
failures, which can lead to Availability, Confidentiality, and
Integrity breaches, in particular if an adversary has control over the
values (e.g., by entering them into a Web form).
The resulting class of attacks, injections (e.g., SQL injections),
is consistently found among the top causes of application security
vulnerabilities and requires particular attention.¶
Where JSONPath is used as a part of a security mechanism, attackers can attempt to provoke unexpected or unpredictable behavior, or take advantage of differences in behavior between JSONPath implementations.¶
Unexpected or unpredictable behavior can arise from an argument with certain constructs described as unpredictable by [RFC8259]. Predictable behavior can be expected, except in relation to the ordering of objects, for any argument conforming with [RFC7493].¶
Other attacks can target the behavior of underlying technologies such as UTF-8 (see Section 10 of [RFC3629]) and the Unicode character set.¶
This appendix is informative.¶
At the time JSONPath was invented, XML was noted for the availability of powerful tools to analyze, transform and selectively extract data from XML documents. [XPath] is one of these tools.¶
In 2007, the need for something solving the same class of problems for the emerging JSON community became apparent, specifically for:¶
(Note that XPath has evolved since 2007, and recent versions even nominally support operating inside JSON values. This appendix only discusses the more widely used version of XPath that was available in 2007.)¶
JSONPath picks up the overall feeling of XPath, but maps the concepts to syntax (and partially semantics) that would be familiar to someone using JSON in a dynamic language.¶
E.g., in popular dynamic programming languages such as JavaScript, Python and PHP, the semantics of the XPath expression¶
/store/book[1]/title¶
can be realized in the expression¶
x.store.book[0].title¶
or, in bracket notation,¶
x['store']['book'][0]['title']¶
with the variable x holding the argument.¶
The JSONPath language was designed to:¶
JSONPath expressions apply to JSON values in the same way
as XPath expressions are used in combination with an XML document.
JSONPath uses $
to refer to the root node of the argument, similar
to XPath's /
at the front.¶
JSONPath expressions move further down the hierarchy using dot notation
($.store.book[0].title
)
or the bracket notation
($['store']['book'][0]['title']
), a lightweight/limited, and a more
heavyweight syntax replacing XPath's /
within query expressions.¶
Both JSONPath and XPath use *
for a wildcard.
The descendant operators, starting with ..
, borrowed from [E4X], are similar to XPath's //
.
The array slicing construct [start:end:step]
is unique to JSONPath,
inspired by [SLICE] from ECMASCRIPT 4.¶
Filter expressions are supported via the syntax ?<boolean expr>
as in¶
$.store.book[?@.price < 10].title¶
Table 21 extends Table 1 by providing a comparison with similar XPath concepts.¶
XPath | JSONPath | Description |
---|---|---|
/
|
$
|
the root XML element |
.
|
@
|
the current XML element |
/
|
. or []
|
child operator |
..
|
n/a | parent operator |
//
|
..name , ..[index] , ..* , or ..[*]
|
descendants (JSONPath borrows this syntax from E4X) |
*
|
*
|
wildcard: All XML elements regardless of their names |
@
|
n/a | attribute access: JSON values do not have attributes |
[]
|
[]
|
subscript operator used to iterate over XML element collections and for predicates |
|
|
[,]
|
Union operator (results in a combination of node sets); called list operator in JSONPath, allows combining member names, array indices, and slices |
n/a |
[start:end:step]
|
array slice operator borrowed from ES4 |
[]
|
?
|
applies a filter (script) expression |
seamless | n/a | expression engine |
()
|
n/a | grouping |
For further illustration, Table 22 shows some XPath expressions and their JSONPath equivalents.¶
XPath | JSONPath | Result |
---|---|---|
/store/book/author
|
$.store.book[*].author
|
the authors of all books in the store |
//author
|
$..author
|
all authors |
/store/*
|
$.store.*
|
all things in store, which are some books and a red bicycle |
/store//price
|
$.store..price
|
the prices of everything in the store |
//book[3]
|
$..book[2]
|
the third book |
//book[last()]
|
$..book[-1]
|
the last book in order |
//book[position()<3]
|
$..book[0,1] $..book[:2]
|
the first two books |
//book[isbn]
|
$..book[?@.isbn]
|
filter all books with isbn number |
//book[price<10]
|
$..book[?@.price<10]
|
filter all books cheaper than 10 |
//*
|
$..*
|
all elements in XML document; all member values and array elements contained in input value |
XPath has a lot more functionality (location paths in unabbreviated syntax, operators and functions) than listed in this comparison. Moreover, there are significant differences in how the subscript operator works in XPath and JSONPath:¶
This appendix is informative.¶
JSONPath is not intended as a replacement for, but as a more powerful companion to, JSON Pointer [RFC6901]. The purposes of the two standards are different.¶
JSON Pointer is for identifying a single value within a JSON value whose structure is known.¶
JSONPath can identify a single value within a JSON value, for example by using a Normalized Path. But JSONPath is also a query syntax that can be used to search for and extract multiple values from JSON values whose structure is known only in a general way.¶
A Normalized JSONPath can be converted into a JSON Pointer by converting the syntax, without knowledge of any JSON value. The inverse is not generally true: a numeric path component in a JSON Pointer may identify a member value of an object or an element of an array. For conversion to a JSONPath query, knowledge of the structure of the JSON value is needed to distinguish these cases.¶
This document is based on Stefan Gössner's original online article defining JSONPath [JSONPath-orig].¶
The books example was taken from http://coli.lili.uni-bielefeld.de/~andreas/Seminare/sommer02/books.xml — a dead link now.¶