The APPROX_COUNT_DISTINCT function counts the approximate number of unique items in a field.
APPROX_COUNT_DISTINCT is only available when your data comes from a BigQuery data source. For other data source types, use COUNT_DISTINCT.Syntax
Parameters
value- a field or expression that contains the items to be counted.
How the APPROX_COUNT_DISTINCT function works
The APPROX_COUNT_DISTINCT function takes one parameter, which can be the name of a metric, dimension, or expression of any type. APPROX_COUNT_DISTINCT returns the approximate number of unique items in that field or expression.
APPROX_COUNT_DISTINCT is more efficient in terms of query processing than COUNT_DISTINCT, but returns less exact results. If your data set is very large, or if the performance of your report is more important than exact counts, consider using APPROX_COUNT_DISTINCT. Using APPROX_COUNT_DISTINCT instead of COUNT_DISTINCT can also help reduce query costs when using BigQuery data sources.
For an in-depth explanation of how approximate aggregation works, see the BigQuery documentation.
Examples
APPROX_COUNT_DISTINCT(Page) - counts the approximate number of unique values in the Page dimension.Limits of APPROX_COUNT_DISTINCT
- The
APPROX_COUNT_DISTINCTfunction is only available when used with BigQuery data sources. - For data sources which do not support
APPROX_COUNT_DISTINCT,APPROX_COUNT_DISTINCTwill act like COUNT_DISTINCT. - You can't apply this function to a pre-aggregated metric (Aggregation type of Auto), or to an expression which is the result of another aggregation function. For example, a formula such as
APPROX_COUNT_DISTINCT(Sessions)in a Google Analytics data source will produce an error.