io.druid.query.groupby.strategy

Class GroupByStrategyV2

java.lang.Object

io.druid.query.groupby.strategy.GroupByStrategyV2

All Implemented Interfaces:

GroupByStrategy

public class GroupByStrategyV2
extends Object
implements GroupByStrategy

Field Summary

Fields

Modifier and Type	Field and Description
static String	CTX_KEY_FUDGE_TIMESTAMP
static String	CTX_KEY_OUTERMOST

Constructor Summary

Constructors

Constructor and Description
GroupByStrategyV2(DruidProcessingConfig processingConfig, com.google.common.base.Supplier<GroupByQueryConfig> configSupplier, StupidPool<ByteBuffer> bufferPool, BlockingPool<ByteBuffer> mergeBufferPool, com.fasterxml.jackson.databind.ObjectMapper spillMapper, QueryWatcher queryWatcher)

Method Summary

Modifier and Type	Method and Description
QueryRunner<Row>	createIntervalChunkingRunner(IntervalChunkingQueryRunnerDecorator decorator, QueryRunner<Row> runner, GroupByQueryQueryToolChest toolChest) Decorate a runner with an interval chunking decorator.
boolean	doMergeResults(GroupByQuery query) Indicates if this query should undergo "mergeResults" or not.
static org.joda.time.DateTime	getUniversalTimestamp(GroupByQuery query) If "query" has a single universal timestamp, return it.
boolean	isCacheable(boolean willMergeRunners) Indicates this strategy is cacheable or not.
Sequence<Row>	mergeResults(QueryRunner<Row> baseRunner, GroupByQuery query, Map<String,Object> responseContext)
QueryRunner<Row>	mergeRunners(com.google.common.util.concurrent.ListeningExecutorService exec, Iterable<QueryRunner<Row>> queryRunners)
GroupByQueryResource	prepareResource(GroupByQuery query, boolean willMergeRunners) Initializes resources required for a broker to process the given query.
Sequence<Row>	process(GroupByQuery query, StorageAdapter storageAdapter)
Sequence<Row>	processSubqueryResult(GroupByQuery subquery, GroupByQuery query, GroupByQueryResource resource, Sequence<Row> subqueryResult)

Methods inherited from class java.lang.Object

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

Field Detail

CTX_KEY_FUDGE_TIMESTAMP

public static final String CTX_KEY_FUDGE_TIMESTAMP

See Also:

Constant Field Values

CTX_KEY_OUTERMOST

public static final String CTX_KEY_OUTERMOST

See Also:

Constant Field Values

Constructor Detail

GroupByStrategyV2

@Inject
public GroupByStrategyV2(DruidProcessingConfig processingConfig,
                                 com.google.common.base.Supplier<GroupByQueryConfig> configSupplier,
                                 StupidPool<ByteBuffer> bufferPool,
                                 BlockingPool<ByteBuffer> mergeBufferPool,
                                 com.fasterxml.jackson.databind.ObjectMapper spillMapper,
                                 QueryWatcher queryWatcher)

Method Detail

getUniversalTimestamp

public static org.joda.time.DateTime getUniversalTimestamp(GroupByQuery query)

If "query" has a single universal timestamp, return it. Otherwise return null. This is useful for keeping timestamps in sync across partial queries that may have different intervals.

Parameters:

query - the query

Returns:

universal timestamp, or null

prepareResource

public GroupByQueryResource prepareResource(GroupByQuery query,
                                            boolean willMergeRunners)

Description copied from interface: GroupByStrategy

Initializes resources required for a broker to process the given query.

Specified by:

prepareResource in interface GroupByStrategy

Parameters:

query - a groupBy query to be processed

Returns:

broker resource

isCacheable

public boolean isCacheable(boolean willMergeRunners)

Description copied from interface: GroupByStrategy

Indicates this strategy is cacheable or not. The willMergeRunners parameter can be used for distinguishing the caller is a broker or a data node.

Specified by:

isCacheable in interface GroupByStrategy

Parameters:

willMergeRunners - indicates that QueryRunnerFactory.mergeRunners(ExecutorService, Iterable) will be called on the cached by-segment results

Returns:

true if this strategy is cacheable, otherwise false.

doMergeResults

public boolean doMergeResults(GroupByQuery query)

Description copied from interface: GroupByStrategy

Indicates if this query should undergo "mergeResults" or not.

Specified by:

doMergeResults in interface GroupByStrategy

createIntervalChunkingRunner

public QueryRunner<Row> createIntervalChunkingRunner(IntervalChunkingQueryRunnerDecorator decorator,
                                                     QueryRunner<Row> runner,
                                                     GroupByQueryQueryToolChest toolChest)

Description copied from interface: GroupByStrategy

Decorate a runner with an interval chunking decorator.

Specified by:

createIntervalChunkingRunner in interface GroupByStrategy

mergeResults

public Sequence<Row> mergeResults(QueryRunner<Row> baseRunner,
                                  GroupByQuery query,
                                  Map<String,Object> responseContext)

Specified by:

mergeResults in interface GroupByStrategy

processSubqueryResult

public Sequence<Row> processSubqueryResult(GroupByQuery subquery,
                                           GroupByQuery query,
                                           GroupByQueryResource resource,
                                           Sequence<Row> subqueryResult)

Specified by:

processSubqueryResult in interface GroupByStrategy

mergeRunners

public QueryRunner<Row> mergeRunners(com.google.common.util.concurrent.ListeningExecutorService exec,
                                     Iterable<QueryRunner<Row>> queryRunners)

Specified by:

mergeRunners in interface GroupByStrategy

process

public Sequence<Row> process(GroupByQuery query,
                             StorageAdapter storageAdapter)

Specified by:

process in interface GroupByStrategy