Package io.trino.plugin.accumulo.index

Class IndexLookup

java.lang.Object

io.trino.plugin.accumulo.index.IndexLookup

public class IndexLookup
extends Object

Class to assist the Trino connector, and maybe external applications, leverage the secondary * index built by the Indexer. Leverages ColumnCardinalityCache to assist in * retrieving row IDs. Currently pretty bound to the Trino connector APIs.

Constructor Summary

Constructors

Constructor	Description
IndexLookup(org.apache.accumulo.core.client.Connector connector, ColumnCardinalityCache cardinalityCache)

Method Summary

Modifier and Type	Method	Description
boolean	applyIndex(String schema, String table, ConnectorSession session, Collection<AccumuloColumnConstraint> constraints, Collection<org.apache.accumulo.core.data.Range> rowIdRanges, List<TabletSplitMetadata> tabletSplits, AccumuloRowSerializer serializer, org.apache.accumulo.core.security.Authorizations auths)	Scans the index table, applying the index based on the given column constraints to return a set of tablet splits.
void	shutdown()

Methods inherited from class java.lang.Object

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

Constructor Detail

IndexLookup

@Inject
public IndexLookup(org.apache.accumulo.core.client.Connector connector,
                   ColumnCardinalityCache cardinalityCache)

Method Detail

shutdown

@PreDestroy
public void shutdown()

applyIndex

public boolean applyIndex(String schema,
                          String table,
                          ConnectorSession session,
                          Collection<AccumuloColumnConstraint> constraints,
                          Collection<org.apache.accumulo.core.data.Range> rowIdRanges,
                          List<TabletSplitMetadata> tabletSplits,
                          AccumuloRowSerializer serializer,
                          org.apache.accumulo.core.security.Authorizations auths)
                   throws Exception

Scans the index table, applying the index based on the given column constraints to return a set of tablet splits.

If this function returns true, the output parameter tabletSplits contains a list of TabletSplitMetadata objects. These in turn contain a collection of Ranges containing the exact row IDs determined using the index.

If this function returns false, the secondary index should not be used. In this case, either the accumulo session has disabled secondary indexing, or the number of row IDs that would be used by the secondary index is greater than the configured threshold (again retrieved from the session).

Parameters:

schema - Schema name

table - Table name

session - Current client session

constraints - All column constraints (this method will filter for if the column is indexed)

rowIdRanges - Collection of Accumulo ranges based on any predicate against a record key

tabletSplits - Output parameter containing the bundles of row IDs determined by the use of the index.

serializer - Instance of a row serializer

auths - Scan-time authorizations

Returns:

True if the tablet splits are valid and should be used, false otherwise

Throws:

Exception - If something bad happens. What are the odds?