com.rpl.rama

Interface Depot

All Superinterfaces:

PartitionedObject

public interface Depot
extends PartitionedObject

Interface for thread-safe depot clients. These are also known as "foreign depots". A depot client is retrieved using a RamaClusterManager on a real cluster or InProcessCluster in a test environment.

Static methods on this class implement built-in depot partitioning schemes for use when declaring a depot in a module with declareDepot.

See Also:

Depots documentation

Nested Class Summary

Nested Classes

Modifier and Type	Interface and Description
static interface	Depot.Declaration Builder-style object to specify options when declaring a depot in a module
static interface	Depot.Partitioning<T> Interface for implementing custom depot partitioners

Method Summary

Modifier and Type	Method and Description
java.util.Map<java.lang.String,java.lang.Object>	append(java.lang.Object data) Append data to depot using AckLevel.ACK.
java.util.Map<java.lang.String,java.lang.Object>	append(java.lang.Object data, AckLevel ackLevel) Append data to depot using specified AckLevel.
java.util.concurrent.CompletableFuture<java.util.Map<java.lang.String,java.lang.Object>>	appendAsync(java.lang.Object data) Append data to depot using AckLevel.ACK.
java.util.concurrent.CompletableFuture<java.util.Map<java.lang.String,java.lang.Object>>	appendAsync(java.lang.Object data, AckLevel ackLevel) Append data to depot using specified AckLevel.
static com.rpl.rama.impl.NativeDepotPartitioning	disallow() Disallows depot clients from appending to this depot.
DepotPartitionInfo	getPartitionInfo(long partitionIndex) Retrieves the start and end offsets of the log for the specified partition index of the depot.
java.util.concurrent.CompletableFuture<DepotPartitionInfo>	getPartitionInfoAsync(long partitionIndex) Retrieves the start and end offsets of the log for the specified partition index of the depot.
static <T extends RamaFunction1> com.rpl.rama.impl.NativeDepotPartitioning	hashBy(java.lang.Class<T> keyExtractorClass) Depot partitioning based on hashing a value from the data being appended.
static com.rpl.rama.impl.NativeDepotPartitioning	hashBy(com.rpl.rama.impl.NativeRamaFunction1 keyExtractor) Depot partitioning based on hashing a value from the data being appended.
static <T extends RamaFunction1> com.rpl.rama.impl.NativeDepotPartitioning	hashBy(java.lang.String key) Depot partitioning based on hashing a value from the data being appended.
static com.rpl.rama.impl.NativeDepotPartitioning	random() Depot partitioning where each append goes to a random partition
java.util.List<java.lang.Object>	read(long partitionIndex, long startOffset, long endOffset) Reads records from the log for the specified partition index of the depot, between the given start and end offsets.
java.util.concurrent.CompletableFuture<java.util.List<java.lang.Object>>	readAsync(long partitionIndex, long startOffset, long endOffset) Reads records from the log for the specified partition index of the depot, between the given start and end offsets.

Methods inherited from interface com.rpl.rama.PartitionedObject

getObjectInfo

Method Detail

random

static com.rpl.rama.impl.NativeDepotPartitioning random()

Depot partitioning where each append goes to a random partition

Returns:

Partitioning scheme implementation

hashBy

static com.rpl.rama.impl.NativeDepotPartitioning hashBy(com.rpl.rama.impl.NativeRamaFunction1 keyExtractor)

Depot partitioning based on hashing a value from the data being appended. This scheme always appends data with the same hashing value to the same depot partition and evenly spreads data with different hashing values across all partitions.

Example: Depot.hashBy(Ops.FIRST) hashes based on first value in appended data

Parameters:

keyExtractor - Built-in function to extract hashing value from appended data

Returns:

Partitioning scheme implementation

hashBy

static <T extends RamaFunction1> com.rpl.rama.impl.NativeDepotPartitioning hashBy(java.lang.Class<T> keyExtractorClass)

Depot partitioning based on hashing a value from the data being appended. This scheme always appends data with the same hashing value to the same depot partition and evenly spreads data with different hashing values across all partitions.

This version uses a custom extractor class. This class must be on the classpath of both the module and client.

Parameters:

keyExtractorClass - Class implementing function to extract hashing value from appended data

Returns:

Partitioning scheme implementation

See Also:

hashBy(NativeRamaFunction1)

hashBy

static <T extends RamaFunction1> com.rpl.rama.impl.NativeDepotPartitioning hashBy(java.lang.String key)

Depot partitioning based on hashing a value from the data being appended. This scheme always appends data with the same hashing value to the same depot partition and evenly spreads data with different hashing values across all partitions.

This version looks up the hashing value by looking up the specified key from the record. The record must implement the Map interface.

Parameters:

key - Key to use to extract the hashing value from the depot record

Returns:

Partitioning scheme implementation

disallow

static com.rpl.rama.impl.NativeDepotPartitioning disallow()

Disallows depot clients from appending to this depot. Appends can still be done from topologies using depotPartitionAppend. Useful for publishing an event stream from a module.

Returns:

Partitioning scheme implementation

append

java.util.Map<java.lang.String,java.lang.Object> append(java.lang.Object data)

Append data to depot using AckLevel.ACK. Blocks until completion.

Parameters:

data - Data to append

Returns:

Map from topology name to streaming ack return for colocated stream topologies

See Also:

Depot documentation

append

java.util.Map<java.lang.String,java.lang.Object> append(java.lang.Object data,
                                                        AckLevel ackLevel)

Append data to depot using specified AckLevel. Blocks until completion.

Parameters:

data - Data to append

ackLevel - Ack level to use

Returns:

Map from topology name to streaming ack return for colocated stream topologies if AckLevel.ACK is set. Otherwise returns empty map.

See Also:

Depot documentation

appendAsync

java.util.concurrent.CompletableFuture<java.util.Map<java.lang.String,java.lang.Object>> appendAsync(java.lang.Object data)

Append data to depot using AckLevel.ACK. Non-blocking.

Parameters:

data - Data to append

Returns:

Future that's delivered with map from topology name to streaming ack return for colocated stream topologies

appendAsync

java.util.concurrent.CompletableFuture<java.util.Map<java.lang.String,java.lang.Object>> appendAsync(java.lang.Object data,
                                                                                                     AckLevel ackLevel)

Append data to depot using specified AckLevel. Non-blocking.

Parameters:

data - Data to append

ackLevel - Ack level to use

Returns:

Future that's delivered once AckLevel is met. If AckLevel.ACK is set, delivered with map from topology name to streaming ack return for colocated stream topologies. Otherwise, delivered map is empty.

getPartitionInfo

DepotPartitionInfo getPartitionInfo(long partitionIndex)

Retrieves the start and end offsets of the log for the specified partition index of the depot. Blocks until completion.

Parameters:

partitionIndex - The partition index for which to retrieve offset information.

Returns:

A DepotPartitionInfo object that provides methods to access the offsets.

getPartitionInfoAsync

java.util.concurrent.CompletableFuture<DepotPartitionInfo> getPartitionInfoAsync(long partitionIndex)

Retrieves the start and end offsets of the log for the specified partition index of the depot. Non-blocking.

Parameters:

partitionIndex - The partition index for which to retrieve offset information.

Returns:

A Future that completes with a DepotPartitionInfo object providing methods to access the offsets.

read

java.util.List<java.lang.Object> read(long partitionIndex,
                                      long startOffset,
                                      long endOffset)

Reads records from the log for the specified partition index of the depot, between the given start and end offsets. Blocks until completion.

Parameters:

partitionIndex - The partition index from which to read records.

startOffset - The starting offset (inclusive) from which to fetch records.

endOffset - The ending offset (exclusive) up to which records will be fetched.

Returns:

A list of records.

readAsync

java.util.concurrent.CompletableFuture<java.util.List<java.lang.Object>> readAsync(long partitionIndex,
                                                                                   long startOffset,
                                                                                   long endOffset)

Reads records from the log for the specified partition index of the depot, between the given start and end offsets. Non-blocking.

Parameters:

partitionIndex - The partition index from which to read records.

startOffset - The starting offset (inclusive) from which to fetch records.

endOffset - The ending offset (exclusive) up to which records will be fetched.

Returns:

A Future that completes with the list of records.