org.apache.spark.sql.columnar

Interface SimpleMetricsCachedBatch

All Superinterfaces:

CachedBatch

public interface SimpleMetricsCachedBatch
extends CachedBatch

A CachedBatch that stores some simple metrics that can be used for filtering of batches with the SimpleMetricsCachedBatchSerializer. The metrics are returned by the stats value. For each column in the batch 5 columns of metadata are needed in the row.

Method Summary

Modifier and Type	Method and Description
long	sizeInBytes()
org.apache.spark.sql.catalyst.InternalRow	stats() Holds stats for each cached column.

Methods inherited from interface org.apache.spark.sql.columnar.CachedBatch

numRows

Method Detail

sizeInBytes

long sizeInBytes()

Specified by:

sizeInBytes in interface CachedBatch

stats

org.apache.spark.sql.catalyst.InternalRow stats()

Holds stats for each cached column. The optional upperBound and lowerBound should be of the same type as the original column. If they are null, then it is assumed that they are not provided, and will not be used for filtering.

• upperBound (optional)
• lowerBound (Optional)
• nullCount: Int
• rowCount: Int
• sizeInBytes: Long

These are repeated for each column in the original cached data.

Returns:

(undocumented)