HuangZhenQiu commented on a change in pull request #6483: [FLINK-7243][flink-formats] Add parquet input format URL: https://github.com/apache/flink/pull/6483#discussion_r225604644
########## File path: flink-formats/flink-parquet/src/main/java/org/apache/flink/formats/parquet/ParquetInputFormat.java ########## @@ -0,0 +1,241 @@ +/* + * 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.flink.formats.parquet; + +import org.apache.flink.api.common.io.CheckpointableInputFormat; +import org.apache.flink.api.common.io.FileInputFormat; +import org.apache.flink.api.common.typeinfo.TypeInformation; +import org.apache.flink.api.java.tuple.Tuple2; +import org.apache.flink.api.java.typeutils.RowTypeInfo; +import org.apache.flink.core.fs.FileInputSplit; +import org.apache.flink.core.fs.Path; +import org.apache.flink.formats.parquet.utils.ParquetRecordReader; +import org.apache.flink.formats.parquet.utils.ParquetSchemaConverter; +import org.apache.flink.formats.parquet.utils.RowReadSupport; +import org.apache.flink.metrics.Counter; +import org.apache.flink.types.Row; +import org.apache.flink.util.Preconditions; + +import org.apache.parquet.ParquetReadOptions; +import org.apache.parquet.filter2.compat.FilterCompat; +import org.apache.parquet.hadoop.ParquetFileReader; +import org.apache.parquet.hadoop.util.HadoopInputFile; +import org.apache.parquet.io.InputFile; +import org.apache.parquet.schema.MessageType; +import org.apache.parquet.schema.Type; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import java.io.IOException; +import java.util.ArrayList; +import java.util.List; + +/** + * The base InputFormat class to read from Parquet files. + * For specific return types the {@link #convert(Row)} method need to be implemented. + * + * <P>Using {@link ParquetRecordReader} to read files instead of {@link org.apache.flink.core.fs.FSDataInputStream}, + * we override {@link #open(FileInputSplit)} and {@link #close()} to change the behaviors. + * + * @param <E> The type of record to read. + */ +public abstract class ParquetInputFormat<E> + extends FileInputFormat<E> + implements CheckpointableInputFormat<FileInputSplit, Tuple2<Long, Long>> { + + private static final long serialVersionUID = 1L; + + private static final Logger LOG = LoggerFactory.getLogger(ParquetInputFormat.class); + + private transient Counter recordConsumed; + + private final TypeInformation[] fieldTypes; + + private final String[] fieldNames; + + private boolean skipThisSplit = false; + + private transient ParquetRecordReader<Row> parquetRecordReader; + + private transient long recordsReadSinceLastSync; + + private long lastSyncedBlock = -1L; + + /** + * Read parquet files with given result parquet schema. + * + * @param path The path of the file to read. + * @param messageType schema of read result + */ + + protected ParquetInputFormat(Path path, MessageType messageType) { + super(path); + RowTypeInfo readType = (RowTypeInfo) ParquetSchemaConverter.fromParquetType(messageType); + this.fieldTypes = readType.getFieldTypes(); + this.fieldNames = readType.getFieldNames(); + // read whole parquet file as one file split + this.unsplittable = true; + } + + /** + * Read parquet files with given result field names and types. + * + * @param path The path of the file to read. + * @param fieldTypes field types of read result of fields + * @param fieldNames field names to read, which can be subset of the parquet schema + */ + protected ParquetInputFormat(Path path, TypeInformation[] fieldTypes, String[] fieldNames) { + super(path); + this.fieldTypes = fieldTypes; + this.fieldNames = fieldNames; + // read whole parquet file as one file split + this.unsplittable = true; + } + + @Override + public Tuple2<Long, Long> getCurrentState() { + return new Tuple2<>(this.lastSyncedBlock, this.recordsReadSinceLastSync); + } + + @Override + public void open(FileInputSplit split) throws IOException { + org.apache.hadoop.conf.Configuration configuration = new org.apache.hadoop.conf.Configuration(); + InputFile inputFile = + HadoopInputFile.fromPath(new org.apache.hadoop.fs.Path(split.getPath().toUri()), configuration); + ParquetReadOptions options = ParquetReadOptions.builder().build(); + ParquetFileReader fileReader = new ParquetFileReader(inputFile, options); + MessageType schema = fileReader.getFileMetaData().getSchema(); + this.skipThisSplit = false; + MessageType readSchema = getReadSchema(schema); + this.parquetRecordReader = new ParquetRecordReader<>(new RowReadSupport(), readSchema, FilterCompat.NOOP); Review comment: Yes. I considered about this. ReadSupported in Parquet interface is bind with converter. It will need a converter for map and pojo type. Given the access method is different for map, row and pojo. It is hard to have generic converter for all of these types. Thus, I decided to convert to row as an intermediate result. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
