org.apache.paimon.mergetree.compact

Class IntervalPartition

java.lang.Object

org.apache.paimon.mergetree.compact.IntervalPartition

public class IntervalPartition
extends Object

Algorithm to partition several data files into the minimum number of SortedRuns.

Constructor Summary

Constructors

Constructor and Description
IntervalPartition(List<DataFileMeta> inputFiles, Comparator<InternalRow> keyComparator)

Method Summary

Modifier and Type	Method and Description
List<List<SortedRun>>	partition() Returns a two-dimensional list of SortedRuns.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

IntervalPartition

public IntervalPartition(List<DataFileMeta> inputFiles,
                         Comparator<InternalRow> keyComparator)

Method Detail

partition

public List<List<SortedRun>> partition()

Returns a two-dimensional list of SortedRuns.

The elements of the outer list are sections. Key intervals between sections do not overlap. This extra layer is to minimize the number of SortedRuns dealt at the same time.

The elements of the inner list are SortedRuns within a section.

Users are expected to use the results by this way:

 for (List<SortedRun> section : algorithm.partition()) {
     // do some merge sorting within section
 }