public class FileStoreLookupFunction extends Object implements Serializable, Closeable
TableFunction
for file store.Modifier and Type | Field and Description |
---|---|
protected org.apache.flink.table.functions.FunctionContext |
functionContext |
Constructor and Description |
---|
FileStoreLookupFunction(Table table,
int[] projection,
int[] joinKeyIndex,
Predicate predicate) |
Modifier and Type | Method and Description |
---|---|
void |
close() |
protected Set<Integer> |
getRequireCachedBucketIds()
Get the set of bucket IDs that need to be cached by the current lookup join subtask.
|
Collection<org.apache.flink.table.data.RowData> |
lookup(org.apache.flink.table.data.RowData keyRow) |
void |
open(org.apache.flink.table.functions.FunctionContext context) |
protected void |
setCacheRowFilter(Filter<InternalRow> cacheRowFilter) |
protected org.apache.flink.table.functions.FunctionContext functionContext
public void open(org.apache.flink.table.functions.FunctionContext context) throws Exception
Exception
public Collection<org.apache.flink.table.data.RowData> lookup(org.apache.flink.table.data.RowData keyRow)
public void close() throws IOException
close
in interface Closeable
close
in interface AutoCloseable
IOException
protected Set<Integer> getRequireCachedBucketIds()
The Flink Planner will distribute data to lookup join nodes based on buckets. This allows paimon to cache only the necessary buckets for each subtask, improving efficiency.
protected void setCacheRowFilter(@Nullable Filter<InternalRow> cacheRowFilter)
Copyright © 2023–2024 The Apache Software Foundation. All rights reserved.