Package io.github.flink.gcp.connector.bigquery.source

Class BigQueryReadClients

java.lang.Object

io.github.flink.gcp.connector.bigquery.source.BigQueryReadClients

@Internal
public final class BigQueryReadClients
extends Object

Storage Read API clients, for the two places the source opens one: the enumerator creates the read session, and every reader opens its assigned streams.

Two factory methods rather than one, because only the reading side has anything to configure: the ReadRows retry budget is inert on a client that only creates sessions, and a caller that had to pass it would be inventing a value. Naming them after the call each serves is what keeps the wrong one from being picked.

public because those two live in sibling packages and Java has no package-tree-internal access — the same reason the connector's metric-name inventory is public.

Method Summary

Modifier and Type	Method	Description
static com.google.cloud.bigquery.storage.v1.BigQueryReadClient	createForReads(String serviceAccountKeyFile, EmulatorEndpoint emulatorEndpoint, int retryMaxAttempts, Runnable onRetry)	Creates the client a reader opens its assigned streams with.
static com.google.cloud.bigquery.storage.v1.BigQueryReadClient	createForSessions(String serviceAccountKeyFile, EmulatorEndpoint emulatorEndpoint)	Creates the client the enumerator creates its read session with.

Methods inherited from class java.lang.Object

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

Method Details

createForSessions

public static com.google.cloud.bigquery.storage.v1.BigQueryReadClient createForSessions(@Nullable
 String serviceAccountKeyFile,
 @Nullable
 EmulatorEndpoint emulatorEndpoint)
                                                                                 throws IOException

Creates the client the enumerator creates its read session with.

CreateReadSession's own retry settings are left as the SDK ships them — a ten-minute budget over DEADLINE_EXCEEDED and UNAVAILABLE. That call happens once per job, on the coordinator thread, before anything has been read, and a failure there is reported immediately rather than sitting inside a fetch.

Parameters:

serviceAccountKeyFile - the service-account key-file path, or null for ADC

emulatorEndpoint - the emulator's gRPC endpoint, or null for BigQuery itself

Returns:

the client; the caller owns it and must close it

Throws:

IOException - if the client cannot be created

createForReads

public static com.google.cloud.bigquery.storage.v1.BigQueryReadClient createForReads(@Nullable
 String serviceAccountKeyFile,
 @Nullable
 EmulatorEndpoint emulatorEndpoint,
 int retryMaxAttempts,
 @Nullable
 Runnable onRetry)
                                                                              throws IOException

Creates the client a reader opens its assigned streams with.

Parameters:

serviceAccountKeyFile - the service-account key-file path, or null for ADC

emulatorEndpoint - the emulator's gRPC endpoint, or null for BigQuery itself

retryMaxAttempts - the bound put on the client's own ReadRows retry

onRetry - run once per retried attempt, or null to observe none

Returns:

the client; the caller owns it and must close it

Throws:

IOException - if the client cannot be created