youngoli commented on a change in pull request #16489: URL: https://github.com/apache/beam/pull/16489#discussion_r798157352
########## File path: sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/bigquery/BigQuerySchemaIOProvider.java ########## @@ -0,0 +1,203 @@ +/* + * 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.bigquery; + +import com.google.api.services.bigquery.model.TableRow; +import com.google.auto.service.AutoService; +import java.io.Serializable; +import java.util.HashMap; +import org.apache.beam.sdk.annotations.Experimental; +import org.apache.beam.sdk.annotations.Internal; +import org.apache.beam.sdk.schemas.Schema; +import org.apache.beam.sdk.schemas.Schema.FieldType; +import org.apache.beam.sdk.schemas.io.SchemaIO; +import org.apache.beam.sdk.schemas.io.SchemaIOProvider; +import org.apache.beam.sdk.schemas.transforms.Convert; +import org.apache.beam.sdk.transforms.PTransform; +import org.apache.beam.sdk.values.PBegin; +import org.apache.beam.sdk.values.PCollection; +import org.apache.beam.sdk.values.PDone; +import org.apache.beam.sdk.values.Row; +import org.checkerframework.checker.nullness.qual.Nullable; + +/** + * An implementation of {@link SchemaIOProvider} for reading and writing to BigQuery with {@link + * BigQueryIO}. For a description of configuration options, see {@link + * BigQuerySchemaIOProvider#configurationSchema()}. + */ +@Internal +@Experimental +@AutoService(SchemaIOProvider.class) +public class BigQuerySchemaIOProvider implements SchemaIOProvider { + + /** Returns an id that uniquely represents this IO. */ + @Override + public String identifier() { + return "bigquery"; + } + + /** + * Returns the expected schema of the configuration object. Note this is distinct from the schema + * of the data source itself. The fields are as follows: + * + * <ul> + * <li>table: Nullable String - Used for reads and writes. Specifies a table to read or write + * to, in the format described in {@link BigQueryHelpers#parseTableSpec}. Used as an input + * to {@link BigQueryIO.TypedRead#from(String)} or {@link BigQueryIO.Write#to(String)}. + * <li>query: Nullable String - Used for reads. Specifies a query to read results from using the + * BigQuery Standard SQL dialect. Used as an input to {@link + * BigQueryIO.TypedRead#fromQuery(String)}. + * <li>queryLocation: Nullable String - Used for reads. Specifies a BigQuery geographic location + * where the query job will be executed. Used as an input to {@link + * BigQueryIO.TypedRead#withQueryLocation(String)}. + * <li>createDisposition: Nullable String - Used for writes. Specifies whether a table should be + * created if it does not exist. Valid inputs are "Never" and "IfNeeded", corresponding to + * values of {@link BigQueryIO.Write.CreateDisposition}. Used as an input to {@link + * BigQueryIO.Write#withCreateDisposition(BigQueryIO.Write.CreateDisposition)}. + * </ul> + */ + @Override + public Schema configurationSchema() { + return Schema.builder() + .addNullableField("table", FieldType.STRING) + .addNullableField("query", FieldType.STRING) + .addNullableField("queryLocation", FieldType.STRING) + .addNullableField("createDisposition", FieldType.STRING) + .build(); + } + + private static final HashMap<String, BigQueryIO.Write.CreateDisposition> createDispositionsMap = + new HashMap<>(); + + static { + createDispositionsMap.put("Never", BigQueryIO.Write.CreateDisposition.CREATE_NEVER); + createDispositionsMap.put("IfNeeded", BigQueryIO.Write.CreateDisposition.CREATE_IF_NEEDED); + } + + /** + * Produces a SchemaIO given a String representing the data's location, the schema of the data + * that resides there, and some IO-specific configuration object. + * + * <p>For BigQuery IO, only the configuration object is used. Location and data schema have no + * effect. Specifying a table and dataset is done through appropriate fields in the configuration + * object, and the data schema is automatically generated from either the input PCollection or + * schema of the BigQuery table. + */ + @Override + public BigQuerySchemaIO from(String location, Row configuration, @Nullable Schema dataSchema) { + return new BigQuerySchemaIO(location, configuration); + } + + /** + * Indicates whether this transform requires a specified data schema. + * + * @return false + */ + @Override + public boolean requiresDataSchema() { + return false; + } + + /** + * Indicates whether the PCollections produced by this transform will contain a bounded or + * unbounded number of elements. + * + * @return Bounded + */ + @Override + public PCollection.IsBounded isBounded() { + return PCollection.IsBounded.BOUNDED; + } + + /** An abstraction to create schema aware IOs. */ + static class BigQuerySchemaIO implements SchemaIO, Serializable { + protected final Row config; + protected final String location; + + BigQuerySchemaIO(String location, Row config) { + this.config = config; + this.location = location; + } + + /** + * Returns the schema of the data, if one was provided. + * + * @return null + */ + @Override + @SuppressWarnings({ + "nullness" // TODO(https://issues.apache.org/jira/browse/BEAM-10402) + }) + public Schema schema() { + return null; + } + + /** Builds a schema aware reader via {@link BigQueryIO#readTableRowsWithSchema()}. */ + @Override + public PTransform<PBegin, PCollection<Row>> buildReader() { + return new PTransform<PBegin, PCollection<Row>>() { + @Override + public PCollection<Row> expand(PBegin input) { + BigQueryIO.TypedRead<TableRow> read = BigQueryIO.readTableRowsWithSchema(); + read = read.withMethod(BigQueryIO.TypedRead.Method.EXPORT); + + String table = config.getString("table"); + if (table != null) { + read = read.from(table); + } + String query = config.getString("query"); + if (query != null) { + read = read.fromQuery(query).usingStandardSql(); + } + String queryLocation = config.getString("queryLocation"); + if (queryLocation != null) { + read = read.withQueryLocation(queryLocation); + } + return input.apply(read).apply(Convert.toRows()); + } + }; + } + + /** Builds a schema aware writer via {@link BigQueryIO#write()}. */ + @Override + public PTransform<PCollection<Row>, PDone> buildWriter() { Review comment: This is probably approximately how we should be doing it. WriteResult is actually a collection of three PCollections as outputs: failedInserts, failedInsertsWithErr, and successfulInserts, with the second one being enabled only if a specific method is called (withExtendedErrorInfo). So basically, it looks like the best approach here would be to add 2-3 outputs to the transform (depending on how many of those PCollections we want to support), and then convert all elements in the original WriteResults from TableRow to Beam Rows, with the built in conversion BigQuery IO currently uses. Doesn't sound too hard, but it does sound non-trivial enough to do in a separate PR. Of course, the assumption there is that adding new outputs to the transform isn't a breaking change to users. Although I suppose in this case, "users" basically means us devs, which isn't that big a deal. I.e. adding WriteResults as outputs would mean modifying the Go and possibly Python wrappers, but otherwise wouldn't impact actual users of Beam... Basically, I'm leaning towards a yes, but we should do this in future PRs. What do you think? -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
