Type: Function
Function: FLATTEN
Syntax: FLATTEN(array_field)
Description: Flatten an array field (e.g. labels) to produce one parent record per value in the array. This operator is useful to turn arrays into single values. Aggregation logic can then be applied on those individual values (e.g. COUNT per individual label).
This operator can only be used in dimension formulas, it cannot be used in metric formulas.
In a filter, FLATTEN does nothing: it returns the array unchanged, so a dimension's formula can be reused in a filter as is. In a filter, CONTAINS(FLATTEN(label_names), "bug") is equivalent to CONTAINS(label_names, "bug"): both keep the records that have the label bug.
Sample usage:
FLATTEN(label_names)will produce one parent record per label in thelabel_namesarray, with each record having a single label.FLATTEN(assignee_usernames)will produce one parent record per user in theassignee_usernamesarray, with each record having a single assignee.Using
FLATTEN(label_names)as dimension andCOUNT()as metric will produce a reporting counting the number of time each label has been used across all records.Using
FLATTEN(label_names)as dimension 1,FLATTEN(assignee_usernames)as dimension 2 andCOUNT()as metric will produce a reporting counting the number of label+assignee pairs.Using
FLATTEN(label_names)as dimension 1,LEFT(DIMENSION(1), 3)as dimension 2 gives the first three letters of each label, next to the label, without writing theFLATTENtwice. SeeDIMENSION.