Overview

The FOR statement allows you to unnest a list into individual rows by expanding the current working table.

<for statement> ::= "FOR" <for item> [ <for ordinality or offset> ]

<for item> ::= 
  <binding variable> "IN" <list value expression>

<for ordinality or offset> ::= 
  "WITH" { "ORDINALITY" | "OFFSET" } <binding variable>

Details

• The name of the binding variable in <for ordinality or offset> must differ from the binding variable in <for item>.
• The <for ordinality or offset> defines a variable to track the position of elements in the list:
  • ORDINALITY provides the ordinal position of each element, starting from 1.
  • OFFSET provides the zero-based position of each element, where the offset position of an element at ordinal position p is p−1.

Unnesting a List

FOR item IN [1,2,3]
RETURN item

Result:

Unnesting a Binding Table

LET user = "Alex"
LET interests = ["tennis", "violin"]
FOR interest IN interests
RETURN user, interest

Result:

Unnesting a Group Variable

This query retrieves one shortest path between C1 and C3, and returns the values of the property label on edges in the path:

MATCH SHORTEST 1 ({_id: "C1"})-[trans:Transfers]-{1,6}({_id: "C3"})
FOR tran IN trans
RETURN tran.label

Result:

In this query, the variable trans, declared in a quantified path, holds a group degree of reference when used outside the path pattern. To convert it to a singleton degree of reference, you can use the FOR statement to unnest it.

WITH ORDINALITY

This query collects elements in even positions from a list to a new list:

FOR item in ["a", "b", "c", "d"] WITH ORDINALITY index // index starts from 1
FILTER index %2 = 0
RETURN collect_list(item)

Result:

WITH OFFSET

This query returns the second element in a list:

FOR item in ["a", "b", "c", "d"] WITH OFFSET index // index starts from 0
FILTER index = 1
RETURN item

Result: