com.datastax.oss.dsbulk.batcher.reactor

Class ReactorStatementBatcher

java.lang.Object

com.datastax.oss.dsbulk.batcher.api.DefaultStatementBatcher

com.datastax.oss.dsbulk.batcher.reactor.ReactorStatementBatcher

All Implemented Interfaces:

ReactiveStatementBatcher, StatementBatcher

public class ReactorStatementBatcher
extends DefaultStatementBatcher
implements ReactiveStatementBatcher

Nested Class Summary

Nested classes/interfaces inherited from class com.datastax.oss.dsbulk.batcher.api.DefaultStatementBatcher

DefaultStatementBatcher.AdaptiveSizingBatchPredicate

Field Summary

Fields inherited from class com.datastax.oss.dsbulk.batcher.api.DefaultStatementBatcher

batchMode, batchType, codecRegistry, maxBatchStatements, maxSizeInBytes, protocolVersion, session

Constructor Summary

Constructors

Constructor and Description
ReactorStatementBatcher() Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the latest stable protocol version and the default CodecRegistry instance.
ReactorStatementBatcher(CqlSession session) Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the given session as its source for the protocol version and the CodecRegistry instance to use.
ReactorStatementBatcher(CqlSession session, BatchMode batchMode) Creates a new ReactorStatementBatcher that produces unlogged batches, operates in the specified batch mode and uses the given session as its source for the protocol version and the CodecRegistry instance to use.
ReactorStatementBatcher(CqlSession session, BatchMode batchMode, BatchType batchType, int maxBatchStatements) Creates a new ReactorStatementBatcher that produces batches of the given batchType, operates in the specified batchMode and uses the given session as its source for the protocol version and the CodecRegistry instance to use.
ReactorStatementBatcher(CqlSession session, BatchMode batchMode, BatchType batchType, int maxBatchStatements, long maxSizeInBytes) Creates a new ReactorStatementBatcher that produces batches of the given batchType, operates in the specified batchMode and uses the given session as its source for the protocol version and the CodecRegistry instance to use.
ReactorStatementBatcher(int maxBatchStatements) Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the latest stable protocol version and the default CodecRegistry instance and the default maximum data size in bytes (unlimited).
ReactorStatementBatcher(int maxBatchStatements, long maxSizeInBytes) Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the latest stable protocol version and the default CodecRegistry instance.
ReactorStatementBatcher(long maxSizeInBytes) Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the latest stable protocol version and the default CodecRegistry instance and the default maximum number of statements (100).

Method Summary

Modifier and Type	Method and Description
Flux<Statement<?>>	batchAll(Publisher<BatchableStatement<?>> statements) Batches together all the given statements into groups of statements, regardless of their grouping key.
Flux<Statement<?>>	batchByGroupingKey(Publisher<BatchableStatement<?>> statements) Batches together the given statements into groups of statements having the same grouping key.

Methods inherited from class com.datastax.oss.dsbulk.batcher.api.DefaultStatementBatcher

batchAll, batchByGroupingKey, groupingKey

Methods inherited from class java.lang.Object

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

Methods inherited from interface com.datastax.oss.dsbulk.batcher.api.StatementBatcher

batchAll, batchAll, batchByGroupingKey, batchByGroupingKey

Constructor Detail

ReactorStatementBatcher

public ReactorStatementBatcher()

Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the latest stable protocol version and the default CodecRegistry instance. It also uses the default maximum number of statements (100) and the default maximum data size in bytes (unlimited).

ReactorStatementBatcher

public ReactorStatementBatcher(int maxBatchStatements)

Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the latest stable protocol version and the default CodecRegistry instance and the default maximum data size in bytes (unlimited). It uses the given maximum number of statements.

Parameters:

maxBatchStatements - The maximum number of statements in a batch. If set to zero or any negative value, the number of statements is considered unlimited.

ReactorStatementBatcher

public ReactorStatementBatcher(long maxSizeInBytes)

Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the latest stable protocol version and the default CodecRegistry instance and the default maximum number of statements (100). It uses the given maximum data size in bytes.

Parameters:

maxSizeInBytes - The maximum number of bytes of data in one batch. If set to zero or any negative value, the data size is considered unlimited.

ReactorStatementBatcher

public ReactorStatementBatcher(int maxBatchStatements,
                               long maxSizeInBytes)

Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the latest stable protocol version and the default CodecRegistry instance. It uses the given maximum number of statements and the maximum data size in bytes.

Parameters:

maxBatchStatements - The maximum number of statements in a batch. If set to zero or any negative value, the number of statements is considered unlimited.

maxSizeInBytes - The maximum number of bytes of data in one batch. If set to zero or any negative value, the data size is considered unlimited.

ReactorStatementBatcher

public ReactorStatementBatcher(@NonNull
                               CqlSession session)

Creates a new ReactorStatementBatcher that produces unlogged batches, operates in partition key mode and uses the given session as its source for the protocol version and the CodecRegistry instance to use. It also uses the default maximum number of statements (100) and the default maximum data size in bytes (unlimited).

Parameters:

session - The session to use; cannot be null.

ReactorStatementBatcher

public ReactorStatementBatcher(@NonNull
                               CqlSession session,
                               @NonNull
                               BatchMode batchMode)

Creates a new ReactorStatementBatcher that produces unlogged batches, operates in the specified batch mode and uses the given session as its source for the protocol version and the CodecRegistry instance to use. It also uses the default maximum number of statements (100) and the default maximum data size in bytes (unlimited).

Parameters:

session - The session to use; cannot be null.

batchMode - The batch mode to use; cannot be null.

ReactorStatementBatcher

public ReactorStatementBatcher(@NonNull
                               CqlSession session,
                               @NonNull
                               BatchMode batchMode,
                               @NonNull
                               BatchType batchType,
                               int maxBatchStatements)

Creates a new ReactorStatementBatcher that produces batches of the given batchType, operates in the specified batchMode and uses the given session as its source for the protocol version and the CodecRegistry instance to use. It uses the given maximum number of statements and the default maximum data size in bytes (unlimited).

Parameters:

session - The session to use; cannot be null.

batchMode - The batch mode to use; cannot be null.

batchType - The batch type to use; cannot be null.

maxBatchStatements - The maximum number of statements in a batch. If set to zero or any negative value, the number of statements is considered unlimited.

ReactorStatementBatcher

public ReactorStatementBatcher(@NonNull
                               CqlSession session,
                               @NonNull
                               BatchMode batchMode,
                               @NonNull
                               BatchType batchType,
                               int maxBatchStatements,
                               long maxSizeInBytes)

Creates a new ReactorStatementBatcher that produces batches of the given batchType, operates in the specified batchMode and uses the given session as its source for the protocol version and the CodecRegistry instance to use. It uses the given maximum number of statements and the maximum data size in bytes.

Parameters:

session - The session to use; cannot be null.

batchMode - The batch mode to use; cannot be null.

batchType - The batch type to use; cannot be null.

maxBatchStatements - The maximum number of statements in a batch. If set to zero or any negative value, the number of statements is considered unlimited.

maxSizeInBytes - The maximum number of bytes of data in one batch. If set to zero or any negative value, the data size is considered unlimited.

Method Detail

batchByGroupingKey

@NonNull
public Flux<Statement<?>> batchByGroupingKey(@NonNull
                                                      Publisher<BatchableStatement<?>> statements)

Description copied from interface: ReactiveStatementBatcher

Batches together the given statements into groups of statements having the same grouping key. Each group size is capped by the maximum number of statements and the maximum data size.

The grouping key to use is determined by the batch mode in use by this statement batcher.

Note that when a resulting group contains only one statement, this method will not create a batch statement containing that single statement; instead, it will return that same statement.

When the number of statements for the same grouping key is greater than the maximum number of statements, or when their total data size is greater than the maximum data size, statements will be split into smaller batches.

When PARTITION_KEY is used, the grouping key is the statement's routing key or routing token, whichever is available.

When REPLICA_SET is used, the grouping key is the replica set owning the statement's routing key or routing token, whichever is available.

Specified by:

batchByGroupingKey in interface ReactiveStatementBatcher

Parameters:

statements - the statements to batch together.

Returns:

A Publisher of batched statements.

batchAll

@NonNull
public Flux<Statement<?>> batchAll(@NonNull
                                            Publisher<BatchableStatement<?>> statements)

Description copied from interface: ReactiveStatementBatcher

Batches together all the given statements into groups of statements, regardless of their grouping key. Each group size is capped by the maximum number of statements and the maximum data size.

Note that when a resulting group contains only one statement, this method will not create a batch statement containing that single statement; instead, it will return that same statement.

When the number of statements for the same grouping key is greater than the maximum number of statements, or when their total data size is greater than the maximum data size, statements will be split into smaller batches.

Use this method with caution; if the given statements do not share the same routing key, the resulting batch could lead to write throughput degradation.

Specified by:

batchAll in interface ReactiveStatementBatcher

Parameters:

statements - the statements to batch together.

Returns:

A Publisher of batched statements.