PhpStorm 2024.2 Help

SQL for MongoDB

In MongoDB, to manipulate data or perform other operations, you write scripts for the mongo shell in JavaScript. For developers who query data primarily with SQL, it might be helpful to use SQL for MongoDB collections as well. PhpStorm will translate your SQL queries to JavaScript.

  • To see the JavaScript version of SQL, right-click a query and select Show JS Script.

  • The current implementation supports SELECT queries and the following сlauses: JOIN, WHERE, GROUP BY, HAVING, ORDER BY, LIMIT, OFFSET.

See the JavaScript translation for an SQL query

  • Right-click a query and select Show JS Script. To copy JavaScript code to the clipboard, click Copy JS Script to Clipboard.

    You can change and run the script from the JS Script Preview.

Data types

PhpStorm supports the following data types:

  • String, Integer, Float, Boolean

  • NULL, NAN, INFINITY

  • Arrays ([1, 2, 3]) and maps ({a: 1, b: 2, c: 3})

Arrays and maps may include any expressions.

SELECT [1, 3 + 1 - 2, price] AS elements FROM sales; SELECT {a: 1 + 2, b: NULL, c: FALSE, d: [1, 2]} FROM sales;

You can use string literals in single or double quotation marks.

Write column names without quotation marks or use a grave accent (`).

SELECT _id, `acquisitions`, `category_code`, `description`, email_address, phone_number FROM companies WHERE category_code = 'social' AND description = "Mobile Dating";

Object constructors

You can use the following object constructors: BinData, HexData, UUID, MD5, ObjectID, Date, ISODate, Code, DBRef, RegExp, NumberInt, NumberLong, NumberDecimal, Timestamp, MinKey, MaxKey.

You may use or skip the new keyword. PhpStorm will add the new keyword to the MongoDB query automatically. This feature does not apply to the Date object constructor because of the difference between Date() new Date(). Date() returns the current date as a string and new Date() as a Date object.

If the function name is unknown, PhpStorm searches for the new keyword before the function name. If the new keyword is present, the name is treated as an object constructor. Otherwise, as a function call. For example, the following query:

SELECT new UnknownTypeConstructor() AS A, new NumberInt() AS B, NumberInt(1) AS C FROM companies;

will be translated to:

db.getSiblingDB("admin").getCollection("companies").aggregate([ { $project: {"A": {$literal: new UnknownTypeConstructor()}, "B": {$literal: new NumberInt()}, "C": {$literal: new NumberInt(1)}, "_id": 0} } ])

Functions

You can use all the available MongoDB aggregation pipeline operators except for map, reduce, filter, and let. Also, you can use COUNT(*) but not COUNT (expression) that is not supported yet.

In MongoDB, some functions require named arguments. To use named arguments, use a PostgreSQL-like syntax for the named parameters (for example, => or :=). You will get an error if you do not specify names for the parametrized functions. Consider the following example for the dateToString function:

SELECT dateToString(date => ISODate(), format := '%H:%M:%S:%L%z', timezone => 'Europe/London') AS 'Date' FROM companies;

You can see the list of all the pipeline operators that support named parameters in Aggregation Pipeline Operators at docs.mongodb.com.

Limitations

  • Due to different approaches to aggregate functions in SQL and MongoDB, you cannot use aggregate functions, such as AVG, SUM, MIN, and MAX, as non-aggregate. For example, SELECT MAX(1,2,3) will not work.

    In MongoDB, functions like AVG, SUM, MIN, and MAX may be aggregate and non-aggregate. They are aggregate if they are placed in the group block. Otherwise, they are non-aggregate.

    In SQL, there is no such dependency, and you can use aggregate functions without the GROUP BY clause (for example, SELECT AVG(x) FROM t).

  • You cannot call aggregate functions within other aggregate functions.

Operators

Currently, the following functionality is supported:

  • Operands: =, ==, >=, <=, >, <, <>, !=, +, -, /, *, %, AND, OR, NOT, LIKE, NOT LIKE, IS, IS NOT, IN, NOT IN, BETWEEN, NOT BETWEEN.

  • You can use wildcards for the LIKE operator. These wildcards are translated into a valid regular expression for MongoDB.

    wildcards for the LIKE operator

SQL clauses

SELECT

The following features are supported for the SELECT queries.

  • Options: DISTINCT, ALL.

  • Aggregate and non-aggregate functions.

  • Access to embedded fields by using dotted names. Consider the following example.

    embedded fields

Limitations

  • Columns may have aliases. Aliases must not contain periods(.).

FROM

  • You can use aliases for tables.

  • You have to use an alias for a nested SELECT statement.

  • You can use nested SELECT queries. For example, the following query is valid:

    nested SELECT queries

Limitations

  • Aliases must not contain periods(.).

JOIN

  • Only JOIN (INNER JOIN) and LEFT JOIN (LEFT OUTER JOIN) are supported.

  • Only the ON condition is supported.

  • Only one condition in the ON part. The following query will result in error:

    one condition in the ON part
  • You can use == or = operators in the ON condition.

  • Multiple JOIN clauses are supported.

Limitations

  • USING is not supported.

  • You cannot use a SELECT statement as the second argument in JOIN clauses.

  • Table names and aliases must not have duplicates in JOIN clauses.

WHERE

  • LIKE and NOT LIKE require a string literal. The following query results in error:

GROUP BY

PhpStorm supports the GROUP BY clause.

The SELECT clause may include expressions that functionally depend on expressions in the GROUP BY clause. You can use embedded fields of a column in the SELECT clause if the column is used in the GROUP BY clause.

GROUP BY

HAVING

PhpStorm supports the HAVING clause.

HAVING

ORDER BY

PhpStorm supports the ORDER BY clause. Instead of ORDER BY, you can use SORT BY.

ORDER BY

LIMIT

PhpStorm supports the LIMIT clause.

LIMIT

OFFSET

PhpStorm supports OFFSET clause.

OFFSET
Last modified: 29 August 2024