This function is intended to be used as part an aggregation to return any single value. When run at scale, there is some randomness to the value that is returned from the aggregated groupings, although randomness in not guaranteed.

In a flat aggregation, in which no aggregate function is applied, it selects the first value that it can retrieve from a column, which is the first value. This function has limited value outside of an aggregation. See Pivot Transform.

Input column might be of Integer, Decimal, String, Object, or Array type. 

Output: Returns a single value from the myRating column.

For more information on the group and limit parameters, see Pivot Transform.

function_col_ref

Name of the column from which to extract a value based on the grouping.

• Literal values are not supported as inputs. 
• Multiple columns and wildcards are not supported.

Example - Aggregating on one customer per month

You want to do some sampling of customer orders on a monthly basis. For your sample, you want to select the sum of orders for one customer each month. 

Source:

Here are the orders for 1Q 2015:

Transformation:

To aggregate this date by month, you must extract the month value from the Date column:

You should now have a new column with three-letter month abbreviations. You can use the following aggregation to gather the sum of one customer's orders for each month:

Results: