Traverse object hierarchies using matching and callbacks.
Install with npm:
$ npm install --save object-scan
const objectScan = require('object-scan');
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' } } };
objectScan(['a.*.f'], { joined: true })(haystack);
// => [ 'a.e.f' ]
- Input traversed exactly once during search
- Dependency free, small in size and very performant
- Separate Object and Array matching
- Wildcard and Regex matching
- Arbitrary depth matching
- Or-clause Syntax
- Exclusion Matching
- Full support for escaping
- Traversal in "delete-safe" order
- Recursion free implementation
- Search syntax validated
- Lots of tests and examples
A needle expression specifies one or more paths to an element (or a set of elements) in a JSON structure. Paths use the dot notation:
store.book[0].title
Rectangular brackets for array path matching.
Examples:
['[2]']
(exact in array)
const haystack = [0, 1, 2, 3, 4];
objectScan(['[2]'], { joined: true })(haystack);
// => [ '[2]' ]
['[1]']
(no match in object)
const haystack = { 0: 'a', 1: 'b', 2: 'c' };
objectScan(['[1]'], { joined: true })(haystack);
// => []
Property name for object property matching.
Examples:
['foo']
(exact in object)
const haystack = { foo: 0, bar: 1 };
objectScan(['foo'], { joined: true })(haystack);
// => [ 'foo' ]
['1']
(no match in array)
const haystack = [0, 1, 2, 3, 4];
objectScan(['1'], { joined: true })(haystack);
// => []
The following characters have special meaning when not escaped:
*
: Match zero or more character+
: Match one or more character?
: Match exactly one character\
: Escape the subsequent character
Wildcards can be used with Array and Object selector.
Examples:
['*']
(top level)
const haystack = { a: { b: 0, c: 1 }, d: 2 };
objectScan(['*'], { joined: true })(haystack);
// => [ 'd', 'a' ]
['[?5]']
(two digit ending in five)
const haystack = [...Array(30).keys()];
objectScan(['[?5]'], { joined: true })(haystack);
// => [ '[25]', '[15]' ]
['a.+.c']
(nested)
const haystack = { a: { b: { c: 0 }, d: { f: 0 } } };
objectScan(['a.+.c'], { joined: true })(haystack);
// => [ 'a.b.c' ]
['a.\\+.c']
(escaped)
const haystack = { a: { b: { c: 0 }, '+': { c: 0 } } };
objectScan(['a.\\+.c'], { joined: true })(haystack);
// => [ 'a.\\+.c' ]
Regex are defined by using parentheses.
Can be used with Array and Object selector.
Examples:
['(^foo)']
(starting with `foo`)
const haystack = { foo: 0, foobar: 1, bar: 2 };
objectScan(['(^foo)'], { joined: true })(haystack);
// => [ 'foobar', 'foo' ]
['[(5)]']
(containing `5`)
const haystack = [...Array(20).keys()];
objectScan(['[(5)]'], { joined: true })(haystack);
// => [ '[15]', '[5]' ]
['[(^[01]$)]']
(`[0]` and `[1]`)
const haystack = ['a', 'b', 'c', 'd'];
objectScan(['[(^[01]$)]'], { joined: true })(haystack);
// => [ '[1]', '[0]' ]
['[(^[^01]$)]']
(other than `[0]` and `[1]`)
const haystack = ['a', 'b', 'c', 'd'];
objectScan(['[(^[^01]$)]'], { joined: true })(haystack);
// => [ '[3]', '[2]' ]
['[*]', '[!(^[01]$)]']
(match all and exclude `[0]` and `[1]`)
const haystack = ['a', 'b', 'c', 'd'];
objectScan(['[*]', '[!(^[01]$)]'], { joined: true })(haystack);
// => [ '[3]', '[2]' ]
There are two types of arbitrary depth matching:
**
: Matches zero or more nestings++
: Matches one or more nestings
Recursions can be combined with a regex by appending the regex.
Examples:
['a.**']
(zero or more nestings under `a`)
const haystack = { a: { b: 0, c: 0 } };
objectScan(['a.**'], { joined: true })(haystack);
// => [ 'a.c', 'a.b', 'a' ]
['a.++']
(one or more nestings under `a`)
const haystack = { a: { b: 0, c: 0 } };
objectScan(['a.++'], { joined: true })(haystack);
// => [ 'a.c', 'a.b' ]
['**(1)']
(all containing `1` at every level)
const haystack = { 1: { 1: ['c', 'd'] }, 510: 'e', foo: { 1: 'f' } };
objectScan(['**(1)'], { joined: true })(haystack);
// => [ '510', '1.1[1]', '1.1', '1' ]
Or Clauses are defined by using curley brackets.
Can be used with Array and Object selector.
Examples:
['[{0,1}]']
(`[0]` and `[1]`)
const haystack = ['a', 'b', 'c', 'd'];
objectScan(['[{0,1}]'], { joined: true })(haystack);
// => [ '[1]', '[0]' ]
['{a,d}.{b,f}']
(`a.b`, `a.f`, `d.b` and `d.f`)
const haystack = { a: { b: 0, c: 1 }, d: { e: 2, f: 3 } };
objectScan(['{a,d}.{b,f}'], { joined: true })(haystack);
// => [ 'd.f', 'a.b' ]
To exclude a path, use exclamation mark.
Examples:
['{a,b},!a']
(only `b`)
const haystack = { a: 0, b: 1 };
objectScan(['{a,b},!a'], {
joined: true,
strict: false
})(haystack);
// => [ 'b' ]
['**,!**.a']
(all except ending in `a`)
const haystack = { a: 0, b: { a: 1, c: 2 } };
objectScan(['**,!**.a'], { joined: true })(haystack);
// => [ 'b.c', 'b' ]
The following characters are considered special and need to
be escaped using \
, if they should be matched in a key:
[
, ]
, {
, }
, (
, )
, ,
, .
, !
, ?
, *
, +
and \
.
Examples:
['\\[1\\]']
(special object key)
const haystack = { '[1]': 0 };
objectScan(['\\[1\\]'], { joined: true })(haystack);
// => [ '\\[1\\]' ]
Signature of all callbacks is
Fn({ key, value, ... })
where:
key
: key that callback is invoked for (respectsjoined
option).value
: value for key.entry
: entry consisting of [key
,value
].property
: current parent property.parent
: current parent.parents
: array of form[parent, grandparent, ...]
.isMatch
: true iff last targeting needle exists and is non-excluding.matchedBy
: all non-excluding needles targeting key.excludedBy
: all excluding needles targeting key.traversedBy
: all needles involved in traversing key.isCircular
: true iffvalue
contained inparents
getKey
: function that returnskey
getValue
: function that returnsvalue
getEntry
: function that returnsentry
getProperty
: function that returnsproperty
getParent
: function that returnsparent
getParents
: function that returnsparents
getIsMatch
: function that returnsisMatch
getMatchedBy
: function that returnsmatchedBy
getExcludedBy
: function that returnsexcludedBy
getTraversedBy
: function that returnstraversedBy
getIsCircular
: function that returnsisCircular
context
: as passed into the search
Notes on Performance:
- Arguments backed by getters use Functions Getter and should be accessed via destructuring to prevent redundant computation.
- Getters should be used to improve performance for conditional access. E.g.
if (isMatch) { getParents() ... }
. - For performance reasons, the same object is passed to all callbacks.
Type: function
Default: undefined
If defined, this callback is invoked for every match. If false
is returned, the current key is excluded from the result.
The return value of this callback has no effect when a search context is provided.
Can be used to do processing as matching keys are traversed.
Invoked in same order as matches would appear in result.
This method is conceptually similar to Array.filter().
Examples:
['**']
(filter function)
const haystack = { a: 0, b: 'bar' };
objectScan(['**'], {
joined: true,
filterFn: ({ value }) => typeof value === 'string'
})(haystack);
// => [ 'b' ]
Type: function
Default: undefined
If defined, this callback is invoked for every key that is traversed by
the search. If true
is returned, all keys nested under the current key are
skipped in the search and from the final result.
Note that breakFn
is invoked before the corresponding filterFn
might be invoked.
Examples:
['**']
(break function)
const haystack = { a: { b: { c: 0 } } };
objectScan(['**'], {
joined: true,
breakFn: ({ key }) => key === 'a.b'
})(haystack);
// => [ 'a.b', 'a' ]
Type: boolean
Default: false
When set to true
the scan immediately returns after the first match.
Examples:
['a', 'b']
(only return first property)
const haystack = { a: 0, b: 1 };
objectScan(['a', 'b'], {
rtn: 'property',
abort: true
})(haystack);
// => 'b'
['[0]', '[1]']
(abort changes count)
const haystack = ['a', 'b'];
objectScan(['[0]', '[1]'], {
rtn: 'count',
abort: true
})(haystack);
// => 1
Type: string
Default: dynamic
Defaults to key
when search context is undefined and to context
otherwise.
Can be explicitly set as:
context
: search context is returnedkey
: matched keys are returnedvalue
: matched values are returnedentry
: matched entries are returnedproperty
: matched properties are returnedparent
: matched parent are returnedparents
: matched parents are returnedbool
: returns true iff a match is foundcount
: returns the match count
When abort is set to true
and the result would be a list, the first match or undefined is returned.
Examples:
['[*]']
(return values)
const haystack = ['a', 'b', 'c'];
objectScan(['[*]'], { rtn: 'value' })(haystack);
// => [ 'c', 'b', 'a' ]
['foo[*]']
(return entries)
const haystack = { foo: ['bar'] };
objectScan(['foo[*]'], { rtn: 'entry' })(haystack);
// => [ [ [ 'foo', 0 ], 'bar' ] ]
['a.b.c', 'a']
(return properties)
const haystack = { a: { b: { c: 0 } } };
objectScan(['a.b.c', 'a'], { rtn: 'property' })(haystack);
// => [ 'c', 'a' ]
['a.b', 'a.c']
(checks for any match, full scan)
const haystack = { a: { b: 0, c: 1 } };
objectScan(['a.b', 'a.c'], { rtn: 'bool' })(haystack);
// => true
['**']
(return not provided context)
const haystack = { a: 0 };
objectScan(['**'], { rtn: 'context' })(haystack);
// => undefined
['a.b.{c,d}']
(return keys with context passed)
const haystack = { a: { b: { c: 0, d: 1 } } };
objectScan(['a.b.{c,d}'], { rtn: 'key' })(haystack, []);
// => [ [ 'a', 'b', 'd' ], [ 'a', 'b', 'c' ] ]
Type: boolean
Default: false
Keys are returned as a string when set to true
instead of as a list.
Setting this option to true
will negatively impact performance.
Note that _.get and _.set fully support lists.
Examples:
['[*]', '[*].foo']
(joined)
const haystack = [0, 1, { foo: 'bar' }];
objectScan(['[*]', '[*].foo'], { joined: true })(haystack);
// => [ '[2].foo', '[2]', '[1]', '[0]' ]
['[*]', '[*].foo']
(not joined)
const haystack = [0, 1, { foo: 'bar' }];
objectScan(['[*]', '[*].foo'])(haystack);
// => [ [ 2, 'foo' ], [ 2 ], [ 1 ], [ 0 ] ]
Type: boolean
Default: true
When set to false
, no array selectors should be used in any needles and arrays are automatically traversed.
Note that the results still include the array selectors.
Examples:
['a', 'b.d']
(automatic array traversal)
const haystack = [{ a: 0 }, { b: [{ c: 1 }, { d: 2 }] }];
objectScan(['a', 'b.d'], {
joined: true,
useArraySelector: false
})(haystack);
// => [ '[1].b[1].d', '[0].a' ]
['']
(top level array matching)
const haystack = [{ a: 0 }, { b: 1 }];
objectScan([''], {
joined: true,
useArraySelector: false
})(haystack);
// => [ '[1]', '[0]' ]
Type: boolean
Default: true
When set to true
, errors are thrown when:
- a path is identical to a previous path
- a path invalidates a previous path
- a path contains consecutive recursions
Examples:
['a.b', 'a.b']
(identical)
const haystack = [];
objectScan(['a.b', 'a.b'], { joined: true })(haystack);
// => 'Error: Redundant Needle Target: "a.b" vs "a.b"'
['a.{b,b}']
(identical, same needle)
const haystack = [];
objectScan(['a.{b,b}'], { joined: true })(haystack);
// => 'Error: Redundant Needle Target: "a.{b,b}" vs "a.{b,b}"'
['a.b', 'a.**']
(invalidates previous)
const haystack = [];
objectScan(['a.b', 'a.**'], { joined: true })(haystack);
// => 'Error: Needle Target Invalidated: "a.b" by "a.**"'
['**.!**']
(consecutive recursion)
const haystack = [];
objectScan(['**.!**'], { joined: true })(haystack);
// => 'Error: Redundant Recursion: "**.!**"'
A context can be passed into a search invocation as a second parameter. It is available in all callbacks and can be used to manage state across a search invocation without having to recompile the search.
By default all matched keys are returned from a search invocation. However, when it is not undefined, the context is returned instead.
Examples:
['**.{c,d,e}']
(sum values)
const haystack = { a: { b: { c: 2, d: 11 }, e: 7 } };
objectScan(['**.{c,d,e}'], {
joined: true,
filterFn: ({ value, context }) => { context.sum += value; }
})(haystack, { sum: 0 });
// => { sum: 20 }
More extensive examples can be found in the tests.
['a.*.f']
(nested)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['a.*.f'], { joined: true })(haystack);
// => [ 'a.e.f' ]
['*.*.*']
(multiple nested)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['*.*.*'], { joined: true })(haystack);
// => [ 'a.e.f', 'a.b.c' ]
['a.*.{c,f}']
(or filter)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['a.*.{c,f}'], { joined: true })(haystack);
// => [ 'a.e.f', 'a.b.c' ]
['a.*.{c,f}']
(or filter, not joined)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['a.*.{c,f}'])(haystack);
// => [ [ 'a', 'e', 'f' ], [ 'a', 'b', 'c' ] ]
['*.*[*]']
(list filter)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['*.*[*]'], { joined: true })(haystack);
// => [ 'a.h[1]', 'a.h[0]' ]
['*[*]']
(list filter, unmatched)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['*[*]'], { joined: true })(haystack);
// => []
['**']
(star recursion)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['**'], { joined: true })(haystack);
// => [ 'k', 'a.h[1]', 'a.h[0]', 'a.h', 'a.e.f', 'a.e', 'a.b.c', 'a.b', 'a' ]
['++.++']
(plus recursion)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['++.++'], { joined: true })(haystack);
// => [ 'a.h[1]', 'a.h[0]', 'a.h', 'a.e.f', 'a.e', 'a.b.c', 'a.b' ]
['**.f']
(star recursion ending in f)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['**.f'], { joined: true })(haystack);
// => [ 'a.e.f' ]
['**[*]']
(star recursion ending in array)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['**[*]'], { joined: true })(haystack);
// => [ 'a.h[1]', 'a.h[0]' ]
['a.*,!a.e']
(exclusion filter)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['a.*,!a.e'], { joined: true })(haystack);
// => [ 'a.h', 'a.b' ]
['**.(^[bc]$)']
(regex matching)
const haystack = { a: { b: { c: 'd' }, e: { f: 'g' }, h: ['i', 'j'] }, k: 'l' };
objectScan(['**.(^[bc]$)'], { joined: true })(haystack);
// => [ 'a.b.c', 'a.b' ]
Top level object(s) are matched by the empty needle ''
. This is useful for matching objects nested in arrays by setting useArraySelector
to false
.
To match the actual empty string as a key, use (^$)
.
Note that the empty string does not work to match top level objects with _.get or _.set.
Examples:
['']
(match top level objects in array)
const haystack = [{}, {}];
objectScan([''], {
joined: true,
useArraySelector: false
})(haystack);
// => [ '[1]', '[0]' ]
['']
(match top level object)
const haystack = {};
objectScan([''], { joined: true })(haystack);
// => [ '' ]
['**.(^$)']
(match empty string keys)
const haystack = { '': 0, a: { '': 1 } };
objectScan(['**.(^$)'])(haystack);
// => [ [ 'a', '' ], [ '' ] ]
Conceptually this package works as follows:
-
During initialization the needles are parsed and built into a search tree. Various information is pre-computed and stored for every node. Finally the search function is returned.
-
When the search function is invoked, the input is traversed simultaneously with the relevant nodes of the search tree. Processing multiple search tree branches in parallel allows for a single traversal of the input.
Having a separate initialization stage allows for a performant search and significant speed ups when applying the same search to different input.