com.bigdata.service.jini.util
Class DumpFederation.SourceDetailRecord

java.lang.Object
  com.bigdata.service.jini.util.DumpFederation.SourceDetailRecord

All Implemented Interfaces:

Serializable

Enclosing class:

DumpFederation

public static class DumpFederation.SourceDetailRecord
extends Object
implements Serializable

A record detailing various things counted on a per-source basis.

Version:

$Id$

Author:

Bryan Thompson

See Also:

Serialized Form

Field Summary

long	entryCount The sum of the entry count for each AbstractBTree in the index partition view.
long	leafCount The sum of the #of leaves in each AbstractBTree in the index partition view.
long	nodeCount The sum of the #of nodes in each AbstractBTree in the index partition view.
long	segmentByteCount The #of bytes across all IndexSegments in the view.
long	segmentLeafByteCount The #of bytes in the leaf region of the IndexSegments in the view.
long	segmentNodeByteCount The #of bytes in the node region of the IndexSegments in the view.

Constructor Summary

DumpFederation.SourceDetailRecord(DumpFederation.SourceDetailRecord[] a)
Ctor returns a record that contains the sum across the given array of record.

DumpFederation.SourceDetailRecord(long entryCount, long nodeCount, long leafCount, long segmentByteCount, long segmentNodeByteCount, long segmentLeafByteCount)

Method Summary

Methods inherited from class java.lang.Object

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

Field Detail

entryCount

public final long entryCount

The sum of the entry count for each AbstractBTree in the index partition view.

Note: This is computed from the IndexSegmentCheckpoint without requiring us to open the IndexSegment itself, so it can be significantly faster if the IndexSegment is trying to fully buffer the nodes region of the file and will also impose less of a memory burden since those node buffers will not be allocated in response to this operation. The only drawback of the sum of the entry counts is that it will overestimate the #of tuples in an index partition after a split until the next compacting merge because historical IndexSegments in the old view will be reused by each of the view generated by the split until the next compacting merge, move, or join.

Note: The exact range count is MUCH too expensive as it requires materializing every tuple in the index partition view!

nodeCount

public final long nodeCount

The sum of the #of nodes in each AbstractBTree in the index partition view.

leafCount

public final long leafCount

The sum of the #of leaves in each AbstractBTree in the index partition view.

segmentByteCount

public final long segmentByteCount

The #of bytes across all IndexSegments in the view.

segmentNodeByteCount

public final long segmentNodeByteCount

The #of bytes in the node region of the IndexSegments in the view.

segmentLeafByteCount

public final long segmentLeafByteCount

The #of bytes in the leaf region of the IndexSegments in the view.

Constructor Detail

DumpFederation.SourceDetailRecord

public DumpFederation.SourceDetailRecord(long entryCount,
                                         long nodeCount,
                                         long leafCount,
                                         long segmentByteCount,
                                         long segmentNodeByteCount,
                                         long segmentLeafByteCount)

Parameters:

entryCount -

segmentByteCount -

segmentNodeByteCount -

segmentLeafByteCount -

DumpFederation.SourceDetailRecord

public DumpFederation.SourceDetailRecord(DumpFederation.SourceDetailRecord[] a)

Ctor returns a record that contains the sum across the given array of record.

Parameters:

a - An array of records to be summed.