org.apache.beam.sdk.io.gcp.bigtable.changestreams

Class UniqueIdGenerator

java.lang.Object

org.apache.beam.sdk.io.gcp.bigtable.changestreams.UniqueIdGenerator

@Internal
public class UniqueIdGenerator
extends java.lang.Object

Generate unique IDs that can be used to differentiate different jobs and partitions.

Constructor Summary

Constructors

Constructor and Description
UniqueIdGenerator()

Method Summary

Modifier and Type	Method and Description
static java.lang.String	generateRowKeyPrefix() Return a random base64 encoded 8 byte string.
static java.lang.String	getNextId() Return a random base64 encoded 8 byte string.

Methods inherited from class java.lang.Object

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

Constructor Detail

UniqueIdGenerator

public UniqueIdGenerator()

Method Detail

getNextId

public static java.lang.String getNextId()

Return a random base64 encoded 8 byte string. This is used to identify streamer of a specific partition. We expect there to be at most single digit duplicates of the same partition at any one time. The odd of collision if we generate <10 values per partition is less than 1 in 10^18.

Returns:

a random 8 byte string.

generateRowKeyPrefix

public static java.lang.String generateRowKeyPrefix()

Return a random base64 encoded 8 byte string. This is used to identify a pipeline. Once a pipeline is complete, the metadata is left there and not cleaned up. It's possible, over the lifetime, there to be thousands and more pipelines in a single metadata table. The odds of collision if we expect 100,000 pipelines is less than 1 in 1 billion.

Returns:

a random 8 byte string.