Package io.trino.plugin.accumulo.index

Class ColumnCardinalityCache

java.lang.Object

io.trino.plugin.accumulo.index.ColumnCardinalityCache

public class ColumnCardinalityCache
extends Object

This class is an indexing utility to cache the cardinality of a column value for every table. Each table has its own cache that is independent of every other, and every column also has its own Guava cache. Use of this utility can have a significant impact for retrieving the cardinality of many columns, preventing unnecessary accesses to the metrics table in Accumulo for a cardinality that won't change much.

Constructor Summary

Constructors

Constructor	Description
ColumnCardinalityCache(org.apache.accumulo.core.client.Connector connector, AccumuloConfig config)

Method Summary

Modifier and Type	Method	Description
com.google.common.collect.Multimap<Long,AccumuloColumnConstraint>	getCardinalities(String schema, String table, org.apache.accumulo.core.security.Authorizations auths, com.google.common.collect.Multimap<AccumuloColumnConstraint,org.apache.accumulo.core.data.Range> idxConstraintRangePairs, long earlyReturnThreshold, io.airlift.units.Duration pollingDuration)	Gets the cardinality for each AccumuloColumnConstraint.
long	getColumnCardinality(String schema, String table, org.apache.accumulo.core.security.Authorizations auths, String family, String qualifier, Collection<org.apache.accumulo.core.data.Range> colValues)	Gets the column cardinality for all of the given range values.
void	shutdown()

Methods inherited from class java.lang.Object

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

Constructor Detail

ColumnCardinalityCache

@Inject
public ColumnCardinalityCache(org.apache.accumulo.core.client.Connector connector,
                              AccumuloConfig config)

Method Detail

shutdown

@PreDestroy
public void shutdown()

getCardinalities

public com.google.common.collect.Multimap<Long,AccumuloColumnConstraint> getCardinalities(String schema,
                                                                                                String table,
                                                                                                org.apache.accumulo.core.security.Authorizations auths,
                                                                                                com.google.common.collect.Multimap<AccumuloColumnConstraint,org.apache.accumulo.core.data.Range> idxConstraintRangePairs,
                                                                                                long earlyReturnThreshold,
                                                                                                io.airlift.units.Duration pollingDuration)

Gets the cardinality for each AccumuloColumnConstraint. Given constraints are expected to be indexed! Who knows what would happen if they weren't!

Parameters:

schema - Schema name

table - Table name

auths - Scan authorizations

idxConstraintRangePairs - Mapping of all ranges for a given constraint

earlyReturnThreshold - Smallest acceptable cardinality to return early while other tasks complete

pollingDuration - Duration for polling the cardinality completion service

Returns:

An immutable multimap of cardinality to column constraint, sorted by cardinality from smallest to largest

Throws:

org.apache.accumulo.core.client.TableNotFoundException - If the metrics table does not exist

ExecutionException - If another error occurs; I really don't even know anymore.

getColumnCardinality

public long getColumnCardinality(String schema,
                                 String table,
                                 org.apache.accumulo.core.security.Authorizations auths,
                                 String family,
                                 String qualifier,
                                 Collection<org.apache.accumulo.core.data.Range> colValues)
                          throws ExecutionException

Gets the column cardinality for all of the given range values. May reach out to the metrics table in Accumulo to retrieve new cache elements.

Parameters:

schema - Table schema

table - Table name

auths - Scan authorizations

family - Accumulo column family

qualifier - Accumulo column qualifier

colValues - All range values to summarize for the cardinality

Returns:

The cardinality of the column

Throws:

ExecutionException