org.apache.hadoop.hbase.client

Interface RawAsyncTable

All Superinterfaces:

AsyncTableBase

@InterfaceAudience.Public
public interface RawAsyncTable
extends AsyncTableBase

A low level asynchronous table.

The implementation is required to be thread safe.

The returned CompletableFuture will be finished directly in the rpc framework's callback thread, so typically you should not do any time consuming work inside these methods, otherwise you will be likely to block at least one connection to RS(even more if the rpc framework uses NIO).

So, only experts that want to build high performance service should use this interface directly, especially for the scan(Scan, RawScanResultConsumer) below.

TODO: For now the only difference between this interface and AsyncTable is the scan method. The RawScanResultConsumer exposes the implementation details of a scan(heartbeat) so it is not suitable for a normal user. If it is still the only difference after we implement most features of AsyncTable, we can think about merge these two interfaces.

Nested Class Summary

Nested Classes

Modifier and Type	Interface and Description
static interface	RawAsyncTable.CoprocessorCallable<S,R> Delegate to a protobuf rpc call.
static interface	RawAsyncTable.CoprocessorCallback<R> The callback when we want to execute a coprocessor call on a range of regions.

Method Summary

Modifier and Type	Method and Description
<S,R> CompletableFuture<R>	coprocessorService(Function<com.google.protobuf.RpcChannel,S> stubMaker, RawAsyncTable.CoprocessorCallable<S,R> callable, byte[] row) Execute the given coprocessor call on the region which contains the given row.
<S,R> void	coprocessorService(Function<com.google.protobuf.RpcChannel,S> stubMaker, RawAsyncTable.CoprocessorCallable<S,R> callable, byte[] startKey, boolean startKeyInclusive, byte[] endKey, boolean endKeyInclusive, RawAsyncTable.CoprocessorCallback<R> callback) Execute the given coprocessor call on the regions which are covered by the range from startKey and endKey.
default <S,R> void	coprocessorService(Function<com.google.protobuf.RpcChannel,S> stubMaker, RawAsyncTable.CoprocessorCallable<S,R> callable, byte[] startKey, byte[] endKey, RawAsyncTable.CoprocessorCallback<R> callback) Execute the given coprocessor call on the regions which are covered by the range from startKey inclusive and endKey exclusive.
void	scan(Scan scan, RawScanResultConsumer consumer) The basic scan API uses the observer pattern.

Methods inherited from interface org.apache.hadoop.hbase.client.AsyncTableBase

append, batch, batchAll, checkAndDelete, checkAndDelete, checkAndMutate, checkAndMutate, checkAndPut, checkAndPut, delete, delete, deleteAll, exists, exists, existsAll, get, get, getAll, getConfiguration, getName, getOperationTimeout, getReadRpcTimeout, getRpcTimeout, getScanTimeout, getWriteRpcTimeout, increment, incrementColumnValue, incrementColumnValue, mutateRow, put, put, putAll, scanAll

Method Detail

scan

void scan(Scan scan,
          RawScanResultConsumer consumer)

The basic scan API uses the observer pattern. All results that match the given scan object will be passed to the given consumer by calling RawScanResultConsumer.onNext. RawScanResultConsumer.onComplete means the scan is finished, and RawScanResultConsumer.onError means we hit an unrecoverable error and the scan is terminated. RawScanResultConsumer.onHeartbeat means the RS is still working but we can not get a valid result to call RawScanResultConsumer.onNext. This is usually because the matched results are too sparse, for example, a filter which almost filters out everything is specified.

Notice that, the methods of the given consumer will be called directly in the rpc framework's callback thread, so typically you should not do any time consuming work inside these methods, otherwise you will be likely to block at least one connection to RS(even more if the rpc framework uses NIO).

Parameters:

scan - A configured Scan object.

consumer - the consumer used to receive results.

coprocessorService

<S,R> CompletableFuture<R> coprocessorService(Function<com.google.protobuf.RpcChannel,S> stubMaker,
                                              RawAsyncTable.CoprocessorCallable<S,R> callable,
                                              byte[] row)

Execute the given coprocessor call on the region which contains the given row.

The stubMaker is just a delegation to the newStub call. Usually it is only a one line lambda expression, like:

 
 channel -> xxxService.newStub(channel)
 
 

Type Parameters:

S - the type of the asynchronous stub

R - the type of the return value

Parameters:

stubMaker - a delegation to the actual newStub call.

callable - a delegation to the actual protobuf rpc call. See the comment of RawAsyncTable.CoprocessorCallable for more details.

row - The row key used to identify the remote region location

Returns:

the return value of the protobuf rpc call, wrapped by a CompletableFuture.

See Also:

RawAsyncTable.CoprocessorCallable

coprocessorService

default <S,R> void coprocessorService(Function<com.google.protobuf.RpcChannel,S> stubMaker,
                                      RawAsyncTable.CoprocessorCallable<S,R> callable,
                                      byte[] startKey,
                                      byte[] endKey,
                                      RawAsyncTable.CoprocessorCallback<R> callback)

Execute the given coprocessor call on the regions which are covered by the range from startKey inclusive and endKey exclusive. See the comment of coprocessorService(Function, CoprocessorCallable, byte[], boolean, byte[], boolean, CoprocessorCallback) for more details.

See Also:

coprocessorService(Function, CoprocessorCallable, byte[], boolean, byte[], boolean, CoprocessorCallback)

coprocessorService

<S,R> void coprocessorService(Function<com.google.protobuf.RpcChannel,S> stubMaker,
                              RawAsyncTable.CoprocessorCallable<S,R> callable,
                              byte[] startKey,
                              boolean startKeyInclusive,
                              byte[] endKey,
                              boolean endKeyInclusive,
                              RawAsyncTable.CoprocessorCallback<R> callback)

Execute the given coprocessor call on the regions which are covered by the range from startKey and endKey. The inclusive of boundaries are specified by startKeyInclusive and endKeyInclusive. The stubMaker is just a delegation to the xxxService.newStub call. Usually it is only a one line lambda expression, like:

 
 channel -> xxxService.newStub(channel)
 
 

Type Parameters:

S - the type of the asynchronous stub

R - the type of the return value

Parameters:

stubMaker - a delegation to the actual newStub call.

callable - a delegation to the actual protobuf rpc call. See the comment of RawAsyncTable.CoprocessorCallable for more details.

startKey - start region selection with region containing this row. If null, the selection will start with the first table region.

startKeyInclusive - whether to include the startKey

endKey - select regions up to and including the region containing this row. If null, selection will continue through the last table region.

endKeyInclusive - whether to include the endKey

callback - callback to get the response. See the comment of RawAsyncTable.CoprocessorCallback for more details.

See Also:

RawAsyncTable.CoprocessorCallable, RawAsyncTable.CoprocessorCallback