remorhaz / php-json-path
JSONPath implementation in PHP
Installs: 107 826
Dependents: 1
Suggesters: 0
Security: 0
Stars: 13
Watchers: 4
Forks: 5
Open Issues: 5
Requires
- php: ~8.1.0 || ~8.2.0 || ~8.3.0
- ext-intl: *
- ext-json: *
- nikic/php-parser: ^4.12 || ^5
- remorhaz/php-json-data: ^0.7
- remorhaz/php-unilex: ^0.5.3
Requires (Dev)
- bamarni/composer-bin-plugin: ^1.8
- phpunit/phpunit: ^10.1 || ^11
README
JSONPath is a simple query language for JSON documents, inspired by XPath for XML and originally designed by Stefan Goessner.
Features
- Accepts encoded JSON strings as well as decoded PHP data as input, supports both representations in output.
- Selects, deletes or replaces parts of JSON document using JSONPath queries.
- Recognizes definite/indefinite JSONPath queries without executing them.
- Transforms indefinite JSONPath query to set of definite queries for given JSON document.
Requirements
- PHP 8
- JSON extension (ext-json) - required by remorhaz/php-json-data that is used to access JSON documents.
- Internationalization functions (ext-intl)
- Tokenizer extension (ext-tokenizer) - required by nikic/php-parser that is used for code generation.
Installation
You can use Composer to install this package:
composer require remorhaz/php-json-path
Usage
Accessing JSON document
You can create accessible JSON document either from encoded JSON string or from decoded JSON data using corresponding node value factory:
use Remorhaz\JSON\Data\Value\EncodedJson; use Remorhaz\JSON\Data\Value\DecodedJson; // Creating document from JSON-encoded string: $encodedValueFactory = EncodedJson\NodeValueFactory::create(); $encodedJson = '{"a":1}'; $document1 = $encodedValueFactory->createValue($encodedJson); // Creating document from decoded JSON data: $decodedValueFactory = DecodedJson\NodeValueFactory::create(); $decodedJson = (object) ['a' => 1]; $document2 = $decodedValueFactory->createValue($decodedJson);
Creating query
You should use query factory to create query from JSONPath expression:
use Remorhaz\JSON\Path\Query\QueryFactory; $queryFactory = QueryFactory::create(); // Creating query that selects all 'a' properties from any document: $query = $queryFactory->createQuery('$..a');
Definite query is the query that defines exactly one path in document. If query includes any filters, wildcards or deep children scan, it is considered indefinite.
Addressable query is the query that returns unprocessed part(s) of the document. If query returns an aggregate function result, it is considered non-addressable.
Processing query
You should use an instance of query processor to execute queries on given JSON documents:
use Remorhaz\JSON\Path\Processor\Processor; $processor = Processor::create();
Selecting part of a JSON document
There are two ways to select part of JSON document using JSONPath query:
- You can get all matching parts in array, using
::select()
method. This works with both definite and indefinite queries. You will get empty array if none of document parts matches your query. - You can get exactly one matching part, using
::selectOne()
method. Note that this works only with definite queries. You will get an exception if your query is indefinite.
use Remorhaz\JSON\Data\Value\EncodedJson; use Remorhaz\JSON\Path\Processor\Processor; use Remorhaz\JSON\Path\Query\QueryFactory; $processor = Processor::create() $queryFactory = QueryFactory::create(); $encodedValueFactory = EncodedJson\NodeValueFactory::create(); $document = $encodedValueFactory->createValue('{"a":{"a":1,"b":2}'); // Selecting all 'a' properties (indefinite query, values exist): $query1 = $queryFactory->createQuery('$..a'); $result1 = $processor->select($query1, $document); var_dump($result1->select()); // array: ['{"a":1,"b":2}', '1'] // Selecting single 'b' property nested in 'a' property (definite query, value exists): $query2 = $queryFactory->createQuery('$.a.b'); $result2 = $processor->selectOne($query2, $document); var_dump($result2->exists()); // boolean: true var_dump($result2->decode()); // integer: 2 // Selecting single 'b' property (definite query, value doesn't exist): $query3 = $queryFactory->createQuery('$.b'); $result3 = $processor->selectOne($query3, $document); var_dump($result3->exists()); // boolean: false var_dump($result3->decode()); // throws exception
Note that you can either encode result(s) of a selection to JSON string(s) or decode them to raw PHP data. Before accessing a result of ::selectOne()
you can check it's existence with ::exists()
method to avoid exception.
Deleting part of a JSON document
To delete part(s) of a JSON document use ::delete()
method. It works only with addressable queries. You will get an exception if your query is non-addressable. If none of document parts match the query you will get the document unchanged. Special case is deleting root of a document - in this case you will get non-existing result.
use Remorhaz\JSON\Data\Value\EncodedJson; use Remorhaz\JSON\Path\Processor\Processor; use Remorhaz\JSON\Path\Query\QueryFactory; $processor = Processor::create() $queryFactory = QueryFactory::create(); $encodedValueFactory = EncodedJson\NodeValueFactory::create(); $document = $encodedValueFactory->createValue('{"a":{"a":1,"b":2}'); // Deleting all 'b' properties (value exists): $query1 = $queryFactory->createQuery('$..b'); $result1 = $processor->delete($query1, $document); var_dump($result1->exists()); // boolean: true var_dump($result1->encode()); // '{"a":{"a":1}}' // Deleting all 'c' properties (value doesn't exist): $query2 = $queryFactory->createQuery('$..c'); $result2 = $processor->delete($query2, $document); var_dump($result1->exists()); // boolean: true var_dump($result1->encode()); // '{"a":{"a":1,"b":2}}' // Deleting root of the document: $query3 = $queryFactory->createValue('$'); $result3 = $processor->delete($query3, $document); var_dump($result3->exists()); // boolean: false var_dump($result3->encode()); // throws exception
Replacing the part of a JSON document with another JSON document
To replace part(s) of a JSON document with another JSON document use ::replace()
method. It works only with addressable queries. You will get an exception if your query is non-addressable. If none of document parts match the query you will get the document unchanged. If the query matches nested parts of a document, you will also get an exception.
use Remorhaz\JSON\Data\Value\EncodedJson; use Remorhaz\JSON\Path\Processor\Processor; use Remorhaz\JSON\Path\Query\QueryFactory; $processor = Processor::create() $queryFactory = QueryFactory::create(); $encodedValueFactory = EncodedJson\NodeValueFactory::create(); $document1 = $encodedValueFactory->createValue('{"a":{"a":1,"b":2}'); $document2 = $encodedValueFactory->createValue('{"c":3}'); // Replacing 'a' property (value exists): $query1 = $queryFactory->createQuery('$.a'); $result1 = $processor->replace($query1, $document1, $document2); var_dump($result1->encode()); // string: '{"a":{"c":3}}' // Replacing all 'c' properties (value doesn't exist) $query2 = $queryFactory->createQuery('$..c'); $result2 = $processor->replace($query2, $document1, $document2); var_dump($result2->encode()); // string: '{"a":{"a":1,"b":2}' // Replacing all 'a' properties (values are nested): $query3 = $queryFactory->createQuery('$..a'); $result3 = $processor->replace($query3, $document1, $document2); // throws exception
Grammar
All JSONPath queries start with abstract $
symbol that denotes outer level object. Internal structure can be
matched with child operators and filters:
Child operators
There are two notations for selecting structure children: dot-notation and bracket-notation.
Dot-notation allows to select either exactly one property or all children (using a wildcard). Double-dot notation walks through the JSON structure recursively.
Bracket-notation allows to select a set of properties/elements:
Aggregate functions
Aggregate functions can be appended to any path in query and it will return calculated value.
The set of aggregate functions and idea itself is taken from Java implementation.
Filter expressions
When filter is being applied to nodeset, it leaves only those nodes for which the expression evaluates to true.
Filter context
Expression @
points to the value to which the filter was applied.
Operators
Comparison operators can be used to compare value with another value or with a literal. Supported operators are:
==
, !=
, >
, >=
, <
and <=
. Brackets can be used for grouping, and logical operators &&
, ||
and !
are also supported. Regular expressions can be matched using =~
operator.
Original definition
Goessner described JSONPath grammar with providing a set of example queries on JSON sample. Here's his original data sample:
{ "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": 19.95 } } }
And here are his original example queries with result descriptions: