BenWhitehead commented on a change in pull request #14261: URL: https://github.com/apache/beam/pull/14261#discussion_r605893374
########## File path: sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/firestore/FirestoreV1.java ########## @@ -0,0 +1,620 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +package org.apache.beam.sdk.io.gcp.firestore; + +import static java.util.Objects.requireNonNull; + +import com.google.firestore.v1.BatchWriteRequest; +import com.google.firestore.v1.WriteResult; +import com.google.rpc.Status; +import java.io.Serializable; +import java.util.List; +import java.util.Objects; +import javax.annotation.concurrent.Immutable; +import org.apache.beam.sdk.io.gcp.firestore.FirestoreV1WriteFn.BatchWriteFnWithDeadLetterQueue; +import org.apache.beam.sdk.io.gcp.firestore.FirestoreV1WriteFn.DefaultBatchWriteFn; +import org.apache.beam.sdk.transforms.PTransform; +import org.apache.beam.sdk.transforms.ParDo; +import org.apache.beam.sdk.transforms.Reshuffle; +import org.apache.beam.sdk.transforms.display.DisplayData; +import org.apache.beam.sdk.transforms.display.HasDisplayData; +import org.apache.beam.sdk.values.PCollection; +import org.apache.beam.sdk.values.PDone; +import org.apache.beam.sdk.values.PInput; +import org.apache.beam.sdk.values.POutput; +import org.checkerframework.checker.nullness.qual.Nullable; + +/** + * {@link FirestoreV1} provides an API which provides lifecycle managed {@link PTransform}s for + * <a target="_blank" rel="noopener noreferrer" href="https://cloud.google.com/firestore/docs/reference/rpc/google.firestore.v1";>Cloud + * Firestore v1 API</a>. + * <p/> + * This class is part of the Firestore Connector DSL and should be accessed via {@link + * FirestoreIO#v1()}. + * <p/> + * All {@link PTransform}s provided by this API use {@link org.apache.beam.sdk.extensions.gcp.options.GcpOptions + * GcpOptions} on {@link org.apache.beam.sdk.options.PipelineOptions PipelineOptions} for + * credentials access and projectId resolution. As such, the lifecycle of gRPC clients and project + * information is scoped to the bundle level, not the worker level. + * <p/> + * + * <h3>Operations</h3> + * <h4>Write</h4> + * To write a {@link PCollection} to Cloud Firestore use {@link FirestoreV1#write()}, picking the + * behavior of the writer. + * + * Writes use Cloud Firestore's BatchWrite api which provides fine grained write semantics. + * + * The default behavior is to fail a bundle if any single write fails with a non-retryable error. + * <pre>{@code + * PCollection<Write> writes = ...; + * PDone sink = writes Review comment: I've read through both of the threads you linked and I'm leaning toward following the `PCollection<TypeDescribingWhatIsComplete>` from the threads. I would Refactor `BatchWrite` to return a `PCollection<NewWriteResultType>` instead of `PDone`. Two different things jump to mind for the `NewWriteResultType` and I'm curious if you think I should pick one over the other? 1. `WriteSuccessSummary` which would contain a simple summary of the `BatchWrite` including the number of writes, and the number of bytes. ```java public final class WriteSuccessSummary { private final int numWrites; private final long numBytes; } ``` 2. `WriteSuccess` which would be emitted for each processed Write and include the `Write`, and the `WriteResult` and `Status` from Firestore ```java public static final class WriteFailure implements Serializable { private final Write write; private final WriteResult writeResult; private final Status status; } ``` I'm personally leaning toward number 1 as it follows the SQL pattern and reduces the amount of data in the pipeline reducing any overhead if it is ignored. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
