GROUPCONCAT

Syntax

GROUPCONCAT(<any>; [<any>])

Description

Creates a list of all unique non-null values seen in a group (currently max. 1000 values are collected). The order of elements isn't stable when running on a cluster, because the elements for each group might be collected on different nodes. If the element order matters, provide an order by expression as a second (optional) argument. This could for example be a timestamp or ID column in your data set.

This is an aggregate function.

Example

Given the following data:

Key	Value	Time_Stamp
group1	a	1
group1	b	4
group1	c	3
group1	d	2
group2	a	4
group2	b	3
group2	c	2
group2	d	1

First create a group, e.g. GROUPBY(#RawData!Key)

Groups
group1
group2

Then use the GROUPCONCAT(#RawData!Value;#RawData!Time_Stamp), and the result is a list ordered by the identifier in relation to the GROUPBY() column.

Groups	Value
group1	[a, d, c, b]
group2	[d, c, b, a]

This function returns a list. If you require a JSON represented as a string wrap this function in TOJSON (E.g., TOJSON(<function>(...)).