Amazon Redshift will no longer support the creation of new Python UDFs starting November 1, 2025.
If you would like to use Python UDFs, create the UDFs prior to that date.
Existing Python UDFs will continue to function as normal. For more information, see the
blog post
HLL_CREATE_SKETCH function
The HLL_CREATE_SKETCH function returns an HLLSKETCH data type that encapsulates the
input expression values. The HLL_CREATE_SKETCH function works with any data type and
ignores NULL values. When there are no rows in a table or all rows are NULL, the
resulting sketch has no index-value pairs such as
{"version":1,"logm":15,"sparse":{"indices":[],"values":[]}}
.
Syntax
HLL_CREATE_SKETCH (aggregate_expression)
Argument
- aggregate_expression
-
Any valid expression that provides the value to an aggregate, such as a column name. NULL values are ignored. This function supports any data type as input except HLLSKETCH, GEOMETRY, GEOGRAPHY, and VARBYTE.
Return type
The HLL_CREATE_SKETCH function returns an HLLSKETCH value.
Examples
The following example returns the HLLSKETCH type for column an_int
in
table a_table
. A JSON object is used to represent a sparse HyperLogLog
sketch when importing, exporting, or printing sketches. A string representation (in
Base64 format) is used to represent a dense HyperLogLog sketch.
CREATE TABLE a_table(an_int INT); INSERT INTO a_table VALUES (1), (2), (3), (4); SELECT hll_create_sketch(an_int) AS sketch FROM a_table; sketch ------------------------------------------------------------------------------------------------------- {"version":1,"logm":15,"sparse":{"indices":[20812342,20850007,22362299,47158030],"values":[1,2,1,1]}} (1 row)