A newer version of this documentation is available.

WITH clause

reference

March 9, 2025

+ 12

Use WITH to create a common table expression. The common table expression may be temporary result set that can be used as a data source for the query, or an expression for later use within a query.

Purpose

Common table expressions or CTEs can be used to simplify complex queries. They can also be particularly useful when a value needs to be used several times in a query.

The WITH clause has comparable functionality to the LET clause. The major difference between the WITH clause and the LET clause is that the WITH clause can come before the SELECT clause, enabling an earlier definition of expressions; whereas the LET clause must come after the FROM clause.

The WITH clause is evaluated once per query block, and LET is evaluated for every object produced by the FROM or JOIN clause.

You can chain WITH clauses. A CTE that you create in one WITH clause may be referenced in a later WITH clause.

Prerequisites

The WITH clause can only be used preceding a SELECT statement, and in order for you to select data from a document or keyspace, you must have the query_select privilege on the document or keyspace. For more details about user roles, see Authorization.

Syntax

ebnf
ViewCopy
with-clause ::= 'WITH' alias 'AS' '(' ( select | expr ) ')'
                 ( ',' alias 'AS' '(' ( select | expr ) ')' )*

Arguments

alias: [Required] String or expression that represents the name of the variable.
select: SELECT statement that returns a temporary result set assigned to alias.
expression: String or expression that represents a value assigned to alias.

Examples

To try the examples in this section, set the query context to the inventory scope in the travel sample dataset. For more information, see Query Context.

Example 1. Use a common table expression to create an expression for use in a query

Find the average number of public likes for each record. Then find all hotels with a greater than average number of public likes.

sql++
Copy
WITH avgLikeCount AS (
  SELECT VALUE AVG(DISTINCT ARRAY_COUNT(cte.public_likes))
  FROM hotel AS cte
)
SELECT hotel.name, ARRAY_COUNT(hotel.public_likes) AS likeCount
FROM hotel
WHERE ARRAY_COUNT(hotel.public_likes) > avgLikeCount[0]
LIMIT 5;

Results

json
Copy
[
  {
    "likeCount": 8,
    "name": "Medway Youth Hostel"
  },
  {
    "likeCount": 7,
    "name": "Le Clos Fleuri"
  },
  {
    "likeCount": 9,
    "name": "Windy Harbour Farm Hotel"
  },
  {
    "likeCount": 5,
    "name": "Avondale Guest House"
  },
  {
    "likeCount": 8,
    "name": "The Bulls Head"
  }
]

Example 2. Use a common table expression to create a record subset for use in a query

Create a recordset of hotel names and their Cleanliness ratings. Then use this recordset to find the names all hotels whose average Cleanliness rating is greater than 4.5.

sql++
Copy
WITH hotels AS (
  SELECT name, reviews[*].ratings[*].Cleanliness
  FROM hotel
)
SELECT hotels.name
FROM hotels
WHERE ARRAY_AVG(hotels.Cleanliness) > 4.5
LIMIT 5;

Results

json
Copy
[
  {
    "name": "The George Hotel"
  },
  {
    "name": "Windy Harbour Farm Hotel"
  },
  {
    "name": "Avondale Guest House"
  },
  {
    "name": "The Bulls Head"
  },
  {
    "name": "Hill House Holiday Cottage"
  }
]