Package org.apache.datasketches.pig.hll

Pig UDFs for HLL sketches.

See: Description

Class Summary

Class	Description
AlgebraicInitial	Class used to calculate the initial pass of an Algebraic sketch operation.
DataToSketch	This is a Pig UDF that builds Sketches from data.
DataToSketchAlgebraicFinal
DataToSketchAlgebraicIntermediate
SketchToEstimate	This is a User Defined Function (UDF) for getting a unique count estimate from an HllSketch
SketchToEstimateAndErrorBounds	This is a User Defined Function (UDF) for obtaining the unique count estimate along with a lower and upper bound from an HllSketch.
SketchToString	This is a User Defined Function (UDF) for "pretty printing" the summary of an HllSketch
UnionSketch	This is a Pig UDF that performs the Union operation on HllSketches.
UnionSketchAlgebraicFinal
UnionSketchAlgebraicIntermediate

Package org.apache.datasketches.pig.hll Description

Pig UDFs for HLL sketches.

These UDFs can be used as a replacement of corresponding Theta sketch UDFs. Notice that intersections and A-not-B operations are not supported by the HLL sketch. Also notice a small difference in the output type of DataToSketch and UnionSketch: HLL sketch UDFs return DataByteArray (BYTEARRAY in Pig), but corresponding Theta sketch UDFs return a Tuple with single DataByteArray inside. This was a historical accident, and we are reluctant to break the compatibility with existing scripts. HLL sketch UDFs don't have to keep this compatibility. As a result, HLL sketch UDFs don't need flatten() around them to remove the Tuple, and internally they don't have to spend extra resources to wrap every output DataByteArray into a Tuple.

Author:

Alexander Saydakov