public static class ClusteringPrefix.Deserializer
extends java.lang.Object
The main reason for this is that when we deserialize rows from sstables, there is many cases where we have a bunch of rows to skip at the beginning of an index block because those rows are before the requested slice. This class make sure we can answer the question "is the next row on disk before the requested slice" with as little work as possible. It does that by providing a comparison method that deserialize only what is needed to decide of the comparison.
Constructor and Description |
---|
Deserializer(ClusteringComparator comparator,
DataInputPlus in,
SerializationHeader header) |
Modifier and Type | Method and Description |
---|---|
int |
compareNextTo(Slice.Bound bound) |
RangeTombstone.Bound |
deserializeNextBound() |
Clustering |
deserializeNextClustering() |
void |
prepare(int flags,
int extendedFlags) |
ClusteringPrefix.Kind |
skipNext() |
public Deserializer(ClusteringComparator comparator, DataInputPlus in, SerializationHeader header)
public void prepare(int flags, int extendedFlags) throws java.io.IOException
java.io.IOException
public int compareNextTo(Slice.Bound bound) throws java.io.IOException
java.io.IOException
public RangeTombstone.Bound deserializeNextBound() throws java.io.IOException
java.io.IOException
public Clustering deserializeNextClustering() throws java.io.IOException
java.io.IOException
public ClusteringPrefix.Kind skipNext() throws java.io.IOException
java.io.IOException
Copyright © 2017 The Apache Software Foundation