The following table describes the required columns of the input table. The input table can have additional columns, but the function ignores them.
|Column Name||Data Type||Description|
|partition_column||INTEGER, BIGINT, NUMERIC, or VARCHAR||Column on which the input data is partitioned. This column must contain all rows that contain the entity to be averaged. For example, if the function is to return the cumulative moving average of a particular stock share price, all transactions of that stock must be in one partition.|
|cma_column||INTEGER, BIGINT, NUMERIC, or DOUBLE PRECISION||Column whose values are to be averaged.|
|order_by_column||INTEGER, BIGINT, TIMESTAMP, TIME||Column by which the input table is ordered.|