tuple_union_theta_double function
Applies to: Databricks Runtime 18.2 and above
Merges a TupleSketch with double summaries and a Theta Sketch using set union.
Syntax
tuple_union_theta_double ( first, second [, lgNomEntries [, mode ]] )
Arguments
- first: A TupleSketch in binary format with double summaries.
- second: A Theta Sketch in binary format, such as from
theta_sketch_aggaggregate function. - lgNomEntries: An optional
INTEGERliteral specifying the log-base-2 of nominal entries. Must be between 4 and 26, inclusive. The default is 12. - mode: An optional
STRINGliteral specifying the aggregation mode. Valid values:'sum','min','max','alwaysone'. The default is'sum'.
Returns
A BINARY value containing the merged TupleSketch with double summaries.
If first or second is NULL, the result is NULL.
Notes
- Argument order is fixed:
firstmust be a TupleSketch andsecondmust be a Theta Sketch. Passing the sketch types in the reverse order raises TUPLE_INVALID_INPUT_SKETCH_BUFFER_FAMILY. - Theta Sketch keys are assigned a default double summary based on
mode:0.0for'sum',+Infinityfor'min',-Infinityfor'max', or1.0for'alwaysone'. - When a key appears in both sketches, summary values are combined according to
mode.
Common error conditions
- TUPLE_INVALID_INPUT_SKETCH_BUFFER
- TUPLE_INVALID_INPUT_SKETCH_BUFFER_FAMILY
- THETA_INVALID_INPUT_SKETCH_BUFFER
- THETA_INVALID_INPUT_SKETCH_BUFFER_FAMILY
- SKETCH_INVALID_LG_NOM_ENTRIES
- DATATYPE_MISMATCH.TYPE_CHECK_FAILURE_WITH_HINT
- TUPLE_INVALID_SKETCH_MODE
Examples
SQL
> SELECT tuple_sketch_estimate_double(
tuple_union_theta_double(
tuple_sketch_agg_double(col1, val1),
theta_sketch_agg(col2)
)
) FROM VALUES (1, 1.0D, 4), (2, 2.0D, 5), (3, 3.0D, 6) tab(col1, val1, col2);
6.0