bvaradar commented on a change in pull request #4910: URL: https://github.com/apache/hudi/pull/4910#discussion_r829984925
########## File path: hudi-client/hudi-spark-client/src/main/java/org/apache/hudi/client/utils/SparkSchemaUtils.java ########## @@ -0,0 +1,499 @@ +/* + * 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.hudi.client.utils; + +import org.apache.hudi.common.util.collection.Pair; +import org.apache.hudi.internal.schema.InternalSchema; +import org.apache.hudi.internal.schema.Type; +import org.apache.hudi.internal.schema.Types; +import org.apache.hudi.internal.schema.utils.InternalSchemaUtils; +import org.apache.spark.sql.execution.vectorized.WritableColumnVector; +import org.apache.spark.sql.types.ArrayType; +import org.apache.spark.sql.types.ArrayType$; +import org.apache.spark.sql.types.BinaryType; +import org.apache.spark.sql.types.BinaryType$; +import org.apache.spark.sql.types.BooleanType; +import org.apache.spark.sql.types.BooleanType$; +import org.apache.spark.sql.types.ByteType; +import org.apache.spark.sql.types.CharType; +import org.apache.spark.sql.types.DataType; +import org.apache.spark.sql.types.DateType; +import org.apache.spark.sql.types.DateType$; +import org.apache.spark.sql.types.Decimal; +import org.apache.spark.sql.types.DecimalType; +import org.apache.spark.sql.types.DecimalType$; +import org.apache.spark.sql.types.DoubleType; +import org.apache.spark.sql.types.DoubleType$; +import org.apache.spark.sql.types.FloatType; +import org.apache.spark.sql.types.FloatType$; +import org.apache.spark.sql.types.IntegerType; +import org.apache.spark.sql.types.IntegerType$; +import org.apache.spark.sql.types.LongType; +import org.apache.spark.sql.types.LongType$; +import org.apache.spark.sql.types.MapType; +import org.apache.spark.sql.types.MapType$; +import org.apache.spark.sql.types.Metadata; +import org.apache.spark.sql.types.ShortType; +import org.apache.spark.sql.types.StringType; +import org.apache.spark.sql.types.StringType$; +import org.apache.spark.sql.types.StructField; +import org.apache.spark.sql.types.StructType; +import org.apache.spark.sql.types.StructType$; +import org.apache.spark.sql.types.TimestampType; +import org.apache.spark.sql.types.TimestampType$; +import org.apache.spark.sql.types.UserDefinedType; +import org.apache.spark.sql.types.VarcharType; + +import java.nio.charset.StandardCharsets; +import java.sql.Date; +import java.util.ArrayList; +import java.util.Deque; +import java.util.LinkedList; +import java.util.List; +import java.util.Map; +import java.util.concurrent.atomic.AtomicInteger; +import java.util.stream.Collectors; + +public class SparkSchemaUtils { Review comment: Can we rename this class to SparkInternalSchemaConverter ########## File path: hudi-client/hudi-client-common/src/main/java/org/apache/hudi/config/HoodieWriteConfig.java ########## @@ -2021,6 +2048,16 @@ public Builder withSchema(String schemaStr) { return this; } + public Builder withInternalSchema(String internalSchemaStr) { Review comment: I did not see any callers for withInternalSchema. where are we planning to use it ? ########## File path: hudi-client/hudi-client-common/src/main/java/org/apache/hudi/table/action/compact/HoodieCompactor.java ########## @@ -145,9 +150,15 @@ public abstract void preCompact( String instantTime, TaskContextSupplier taskContextSupplier) throws IOException { FileSystem fs = metaClient.getFs(); - - Schema readerSchema = HoodieAvroUtils.addMetadataFields( - new Schema.Parser().parse(config.getSchema()), config.allowOperationMetadataField()); + Schema readerSchema; + Option<InternalSchema> internalSchemaOption = Option.empty(); + if (!StringUtils.isNullOrEmpty(config.getInternalSchema())) { + readerSchema = new Schema.Parser().parse(config.getSchema()); Review comment: Don't we need to add hoodie metadata fields just like in the else section ? ########## File path: hudi-common/src/main/java/org/apache/hudi/common/table/log/AbstractHoodieLogRecordReader.java ########## @@ -361,15 +372,38 @@ private boolean isNewInstantBlock(HoodieLogBlock logBlock) { */ private void processDataBlock(HoodieDataBlock dataBlock, Option<List<String>> keys) throws Exception { try (ClosableIterator<IndexedRecord> recordItr = dataBlock.getRecordItr(keys.orElse(Collections.emptyList()))) { + boolean headRecord = true; + Option<Schema> schemaOption = Option.empty(); while (recordItr.hasNext()) { - IndexedRecord record = recordItr.next(); + IndexedRecord currentRecord = recordItr.next(); + if (headRecord) { + schemaOption = getMergedSchema(dataBlock, currentRecord); + headRecord = false; + } + IndexedRecord record = schemaOption.isPresent() ? AvroSchemaUtil.rewriteRecord(currentRecord, schemaOption.get()) : currentRecord; processNextRecord(createHoodieRecord(record, this.hoodieTableMetaClient.getTableConfig(), this.payloadClassFQN, this.preCombineField, this.withOperationField, this.simpleKeyGenFields, this.partitionName)); totalLogRecords.incrementAndGet(); } } } + private Option<Schema> getMergedSchema(HoodieDataBlock dataBlock, IndexedRecord record) { Review comment: This is an important place where we can add descriptive java-docs around this method . ########## File path: hudi-common/src/main/java/org/apache/hudi/common/table/HoodieTableMetaClient.java ########## @@ -89,6 +89,8 @@ public static final String BOOTSTRAP_INDEX_BY_FILE_ID_FOLDER_PATH = BOOTSTRAP_INDEX_ROOT_FOLDER_PATH + Path.SEPARATOR + ".fileids"; + public static final String SCHEMA_FOLDER_NAME = ".schema"; Review comment: We need to create this directory explicitly (for FileSystems that require explicit mkdir) in HoodieTableMetaClient.initTableAndGetMetaClient ########## File path: hudi-client/hudi-spark-client/src/main/java/org/apache/hudi/table/action/commit/SparkMergeHelper.java ########## @@ -77,14 +89,39 @@ public void runMerge(HoodieTable<T, JavaRDD<HoodieRecord<T>>, JavaRDD<HoodieKey> readSchema = mergeHandle.getWriterSchemaWithMetaFields(); } + Option<InternalSchema> querySchemaOpt = SerDeHelper.fromJson(table.getConfig().getInternalSchema()); + Boolean needToReWriteRecord = false; + // to do support bootstrap + if (querySchemaOpt.isPresent() && !baseFile.getBootstrapBaseFile().isPresent()) { + // check implicitly add columns, and position reorder(spark sql may change cols order) + InternalSchema querySchema = AvroSchemaUtil.evolutionSchemaFromNewAvroSchema(readSchema, querySchemaOpt.get(), true); Review comment: Thanks ########## File path: hudi-client/hudi-spark-client/src/main/scala/org/apache/hudi/HoodieSparkUtils.scala ########## @@ -297,17 +303,25 @@ object HoodieSparkUtils extends SparkAdapterSupport { AttributeReference(columnName, field.get.dataType, field.get.nullable)() } - def getRequiredSchema(tableAvroSchema: Schema, requiredColumns: Array[String]): (Schema, StructType) = { - // First get the required avro-schema, then convert the avro-schema to spark schema. - val name2Fields = tableAvroSchema.getFields.asScala.map(f => f.name() -> f).toMap - // Here have to create a new Schema.Field object - // to prevent throwing exceptions like "org.apache.avro.AvroRuntimeException: Field already used". - val requiredFields = requiredColumns.map(c => name2Fields(c)) - .map(f => new Schema.Field(f.name(), f.schema(), f.doc(), f.defaultVal(), f.order())).toList - val requiredAvroSchema = Schema.createRecord(tableAvroSchema.getName, tableAvroSchema.getDoc, - tableAvroSchema.getNamespace, tableAvroSchema.isError, requiredFields.asJava) - val requiredStructSchema = AvroConversionUtils.convertAvroSchemaToStructType(requiredAvroSchema) - (requiredAvroSchema, requiredStructSchema) + def getRequiredSchema(tableAvroSchema: Schema, requiredColumns: Array[String], internalSchema: InternalSchema = InternalSchema.getDummyInternalSchema): (Schema, StructType, InternalSchema) = { + if (internalSchema.isDummySchema) { + // First get the required avro-schema, then convert the avro-schema to spark schema. + val name2Fields = tableAvroSchema.getFields.asScala.map(f => f.name() -> f).toMap + // Here have to create a new Schema.Field object + // to prevent throwing exceptions like "org.apache.avro.AvroRuntimeException: Field already used". + val requiredFields = requiredColumns.map(c => name2Fields(c)) + .map(f => new Schema.Field(f.name(), f.schema(), f.doc(), f.defaultVal(), f.order())).toList + val requiredAvroSchema = Schema.createRecord(tableAvroSchema.getName, tableAvroSchema.getDoc, + tableAvroSchema.getNamespace, tableAvroSchema.isError, requiredFields.asJava) + val requiredStructSchema = AvroConversionUtils.convertAvroSchemaToStructType(requiredAvroSchema) + (requiredAvroSchema, requiredStructSchema, internalSchema) + } else { + // support nested project Review comment: Question: Is this a TODO or just a comment saying that nested projection is supported. ########## File path: hudi-client/hudi-spark-client/src/main/java/org/apache/hudi/table/action/commit/SparkMergeHelper.java ########## @@ -77,14 +89,39 @@ public void runMerge(HoodieTable<T, JavaRDD<HoodieRecord<T>>, JavaRDD<HoodieKey> readSchema = mergeHandle.getWriterSchemaWithMetaFields(); } + Option<InternalSchema> querySchemaOpt = SerDeHelper.fromJson(table.getConfig().getInternalSchema()); + Boolean needToReWriteRecord = false; + // to do support bootstrap + if (querySchemaOpt.isPresent() && !baseFile.getBootstrapBaseFile().isPresent()) { Review comment: Should we also check for querySchema not same as dummy schema here ? ########## File path: hudi-client/hudi-spark-client/src/main/java/org/apache/hudi/client/utils/SparkSchemaUtils.java ########## @@ -0,0 +1,499 @@ +/* + * 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.hudi.client.utils; + +import org.apache.hudi.common.util.collection.Pair; +import org.apache.hudi.internal.schema.InternalSchema; +import org.apache.hudi.internal.schema.Type; +import org.apache.hudi.internal.schema.Types; +import org.apache.hudi.internal.schema.utils.InternalSchemaUtils; +import org.apache.spark.sql.execution.vectorized.WritableColumnVector; +import org.apache.spark.sql.types.ArrayType; +import org.apache.spark.sql.types.ArrayType$; +import org.apache.spark.sql.types.BinaryType; +import org.apache.spark.sql.types.BinaryType$; +import org.apache.spark.sql.types.BooleanType; +import org.apache.spark.sql.types.BooleanType$; +import org.apache.spark.sql.types.ByteType; +import org.apache.spark.sql.types.CharType; +import org.apache.spark.sql.types.DataType; +import org.apache.spark.sql.types.DateType; +import org.apache.spark.sql.types.DateType$; +import org.apache.spark.sql.types.Decimal; +import org.apache.spark.sql.types.DecimalType; +import org.apache.spark.sql.types.DecimalType$; +import org.apache.spark.sql.types.DoubleType; +import org.apache.spark.sql.types.DoubleType$; +import org.apache.spark.sql.types.FloatType; +import org.apache.spark.sql.types.FloatType$; +import org.apache.spark.sql.types.IntegerType; +import org.apache.spark.sql.types.IntegerType$; +import org.apache.spark.sql.types.LongType; +import org.apache.spark.sql.types.LongType$; +import org.apache.spark.sql.types.MapType; +import org.apache.spark.sql.types.MapType$; +import org.apache.spark.sql.types.Metadata; +import org.apache.spark.sql.types.ShortType; +import org.apache.spark.sql.types.StringType; +import org.apache.spark.sql.types.StringType$; +import org.apache.spark.sql.types.StructField; +import org.apache.spark.sql.types.StructType; +import org.apache.spark.sql.types.StructType$; +import org.apache.spark.sql.types.TimestampType; +import org.apache.spark.sql.types.TimestampType$; +import org.apache.spark.sql.types.UserDefinedType; +import org.apache.spark.sql.types.VarcharType; + +import java.nio.charset.StandardCharsets; +import java.sql.Date; +import java.util.ArrayList; +import java.util.Deque; +import java.util.LinkedList; +import java.util.List; +import java.util.Map; +import java.util.concurrent.atomic.AtomicInteger; +import java.util.stream.Collectors; + +public class SparkSchemaUtils { + private SparkSchemaUtils() { + + } + + public static final String HOODIE_QUERY_SCHEMA = "hadoop.hoodie.querySchema"; Review comment: Any specific reasons why the configuration name starts with hadoop. instead of hoodie. It looks like we are using this to pass internal schema along. In that case, can we rename as hoodie.schema.internal.querySchema ? Similar question for hadoop.hoodie.tablePath ########## File path: hudi-common/src/main/java/org/apache/hudi/common/table/log/AbstractHoodieLogRecordReader.java ########## @@ -361,15 +372,38 @@ private boolean isNewInstantBlock(HoodieLogBlock logBlock) { */ private void processDataBlock(HoodieDataBlock dataBlock, Option<List<String>> keys) throws Exception { try (ClosableIterator<IndexedRecord> recordItr = dataBlock.getRecordItr(keys.orElse(Collections.emptyList()))) { + boolean headRecord = true; Review comment: Hoodie data log block stores schema in its header. So, Instead of reading the schema from the first record, can you directly use the header schema ? You can look at HoodieDataBlock.getWriterSchema(super.getLogBlockHeader()) for some idea ########## File path: hudi-client/hudi-spark-client/src/main/java/org/apache/hudi/client/utils/SparkSchemaUtils.java ########## @@ -0,0 +1,499 @@ +/* + * 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.hudi.client.utils; + +import org.apache.hudi.common.util.collection.Pair; +import org.apache.hudi.internal.schema.InternalSchema; +import org.apache.hudi.internal.schema.Type; +import org.apache.hudi.internal.schema.Types; +import org.apache.hudi.internal.schema.utils.InternalSchemaUtils; +import org.apache.spark.sql.execution.vectorized.WritableColumnVector; +import org.apache.spark.sql.types.ArrayType; +import org.apache.spark.sql.types.ArrayType$; +import org.apache.spark.sql.types.BinaryType; +import org.apache.spark.sql.types.BinaryType$; +import org.apache.spark.sql.types.BooleanType; +import org.apache.spark.sql.types.BooleanType$; +import org.apache.spark.sql.types.ByteType; +import org.apache.spark.sql.types.CharType; +import org.apache.spark.sql.types.DataType; +import org.apache.spark.sql.types.DateType; +import org.apache.spark.sql.types.DateType$; +import org.apache.spark.sql.types.Decimal; +import org.apache.spark.sql.types.DecimalType; +import org.apache.spark.sql.types.DecimalType$; +import org.apache.spark.sql.types.DoubleType; +import org.apache.spark.sql.types.DoubleType$; +import org.apache.spark.sql.types.FloatType; +import org.apache.spark.sql.types.FloatType$; +import org.apache.spark.sql.types.IntegerType; +import org.apache.spark.sql.types.IntegerType$; +import org.apache.spark.sql.types.LongType; +import org.apache.spark.sql.types.LongType$; +import org.apache.spark.sql.types.MapType; +import org.apache.spark.sql.types.MapType$; +import org.apache.spark.sql.types.Metadata; +import org.apache.spark.sql.types.ShortType; +import org.apache.spark.sql.types.StringType; +import org.apache.spark.sql.types.StringType$; +import org.apache.spark.sql.types.StructField; +import org.apache.spark.sql.types.StructType; +import org.apache.spark.sql.types.StructType$; +import org.apache.spark.sql.types.TimestampType; +import org.apache.spark.sql.types.TimestampType$; +import org.apache.spark.sql.types.UserDefinedType; +import org.apache.spark.sql.types.VarcharType; + +import java.nio.charset.StandardCharsets; +import java.sql.Date; +import java.util.ArrayList; +import java.util.Deque; +import java.util.LinkedList; +import java.util.List; +import java.util.Map; +import java.util.concurrent.atomic.AtomicInteger; +import java.util.stream.Collectors; + +public class SparkSchemaUtils { + private SparkSchemaUtils() { + + } + + public static final String HOODIE_QUERY_SCHEMA = "hadoop.hoodie.querySchema"; + public static final String HOODIE_TABLE_PATH = "hadoop.hoodie.tablePath"; Review comment: same question -- 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]
