[GitHub] [flink] leonardBang commented on a change in pull request #11755: [FLINK-14257][Connectors / FileSystem]Integrate csv to FileSystemTableFactory

[GitBox <gi...@apache.org>]

leonardBang commented on a change in pull request #11755:
URL: https://github.com/apache/flink/pull/11755#discussion_r422984520



##########
File path: flink-formats/flink-csv/src/main/java/org/apache/flink/formats/csv/CsvRowDataFileSystemFormatFactory.java
##########
@@ -0,0 +1,352 @@
+package org.apache.flink.formats.csv;
+
+import org.apache.flink.api.common.io.InputFormat;
+import org.apache.flink.api.common.serialization.BulkWriter;
+import org.apache.flink.api.common.serialization.Encoder;
+import org.apache.flink.api.java.typeutils.GenericTypeInfo;
+import org.apache.flink.core.fs.FileInputSplit;
+import org.apache.flink.core.fs.Path;
+import org.apache.flink.formats.csv.CsvRowDataDeserializationSchema.DeserializationRuntimeConverter;
+import org.apache.flink.table.data.GenericRowData;
+import org.apache.flink.table.data.RowData;
+import org.apache.flink.table.descriptors.DescriptorProperties;
+import org.apache.flink.table.factories.FileSystemFormatFactory;
+import org.apache.flink.table.types.DataType;
+import org.apache.flink.table.types.logical.RowType;
+import org.apache.flink.table.utils.PartitionPathUtils;
+
+import org.apache.flink.shaded.jackson2.com.fasterxml.jackson.databind.JsonNode;
+import org.apache.flink.shaded.jackson2.com.fasterxml.jackson.databind.MappingIterator;
+import org.apache.flink.shaded.jackson2.com.fasterxml.jackson.dataformat.csv.CsvMapper;
+import org.apache.flink.shaded.jackson2.com.fasterxml.jackson.dataformat.csv.CsvSchema;
+
+import java.io.BufferedReader;
+import java.io.IOException;
+import java.io.InputStreamReader;
+import java.io.OutputStream;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+import java.util.NoSuchElementException;
+import java.util.Optional;
+import java.util.stream.Collectors;
+
+import static org.apache.flink.table.descriptors.CsvValidator.FORMAT_ALLOW_COMMENTS;
+import static org.apache.flink.table.descriptors.CsvValidator.FORMAT_ARRAY_ELEMENT_DELIMITER;
+import static org.apache.flink.table.descriptors.CsvValidator.FORMAT_DISABLE_QUOTE_CHARACTER;
+import static org.apache.flink.table.descriptors.CsvValidator.FORMAT_ESCAPE_CHARACTER;
+import static org.apache.flink.table.descriptors.CsvValidator.FORMAT_FIELD_DELIMITER;
+import static org.apache.flink.table.descriptors.CsvValidator.FORMAT_IGNORE_PARSE_ERRORS;
+import static org.apache.flink.table.descriptors.CsvValidator.FORMAT_LINE_DELIMITER;
+import static org.apache.flink.table.descriptors.CsvValidator.FORMAT_NULL_LITERAL;
+import static org.apache.flink.table.descriptors.CsvValidator.FORMAT_QUOTE_CHARACTER;
+import static org.apache.flink.table.descriptors.FormatDescriptorValidator.FORMAT;
+
+/**
+ * CSV format factory for file system.
+ */
+public class CsvRowDataFileSystemFormatFactory implements FileSystemFormatFactory {
+	@Override
+	public InputFormat<RowData, ?> createReader(ReaderContext context) {
+
+		DescriptorProperties properties = getValidatedProperties(context.getFormatProperties());
+
+		RowType formatRowType = context.getFormatRowType();
+
+		String[] fieldNames = context.getSchema().getFieldNames();
+		List<String> projectFields = Arrays.stream(context.getProjectFields())
+			.mapToObj(idx -> fieldNames[idx])
+			.collect(Collectors.toList());
+		List<String> csvFields = Arrays.stream(fieldNames)
+			.filter(field -> !context.getPartitionKeys().contains(field))
+			.collect(Collectors.toList());
+
+		int[] csvSelectFieldToProjectFieldMapping = context.getFormatProjectFields().stream()
+			.mapToInt(projectFields::indexOf)
+			.toArray();
+		int[] csvSelectFieldToCsvFieldMapping = context.getFormatProjectFields().stream()
+			.mapToInt(csvFields::indexOf)
+			.toArray();
+
+		CsvSchema csvSchema = buildCsvSchema(formatRowType, properties);
+
+		boolean ignoreParseErrors = properties.getOptionalBoolean(FORMAT_IGNORE_PARSE_ERRORS).orElse(false);
+
+		return new CsvInputFormat(
+			context.getPaths(),
+			context.getSchema().getFieldDataTypes(),
+			context.getSchema().getFieldNames(),
+			csvSchema,
+			formatRowType,
+			context.getProjectFields(),
+			context.getPartitionKeys(),
+			context.getDefaultPartName(),
+			context.getPushedDownLimit(),
+			csvSelectFieldToProjectFieldMapping,
+			csvSelectFieldToCsvFieldMapping,
+			ignoreParseErrors);
+	}
+
+	private CsvSchema buildCsvSchema(RowType rowType, DescriptorProperties properties) {
+		CsvSchema csvSchema = CsvRowSchemaConverter.convert(rowType);
+		//format properties
+		if (properties.containsKey(FORMAT_FIELD_DELIMITER)) {
+			csvSchema = csvSchema.rebuild()
+				.setColumnSeparator(properties.getCharacter(FORMAT_FIELD_DELIMITER))
+				.build();
+		}
+		if (properties.containsKey(FORMAT_QUOTE_CHARACTER)) {
+			csvSchema = csvSchema.rebuild()
+				.setQuoteChar(properties.getCharacter(FORMAT_QUOTE_CHARACTER))
+				.build();
+		}
+		if (properties.containsKey(FORMAT_ALLOW_COMMENTS)) {
+			csvSchema = csvSchema.rebuild()
+				.setAllowComments(properties.getBoolean(FORMAT_ALLOW_COMMENTS))
+				.build();
+		}
+		if (properties.containsKey(FORMAT_ARRAY_ELEMENT_DELIMITER)) {
+			csvSchema = csvSchema.rebuild()
+				.setArrayElementSeparator(properties.getString(FORMAT_ARRAY_ELEMENT_DELIMITER))
+				.build();
+		}
+		if (properties.containsKey(FORMAT_ESCAPE_CHARACTER)) {
+			csvSchema = csvSchema.rebuild()
+				.setEscapeChar(properties.getCharacter(FORMAT_ESCAPE_CHARACTER))
+				.build();
+		}
+		if (properties.containsKey(FORMAT_NULL_LITERAL)) {
+			csvSchema = csvSchema.rebuild()
+				.setNullValue(properties.getString(FORMAT_NULL_LITERAL))
+				.build();
+		}
+		if (properties.containsKey(FORMAT_LINE_DELIMITER)) {
+			csvSchema = csvSchema.rebuild()
+				.setLineSeparator(properties.getString(FORMAT_LINE_DELIMITER))
+				.build();
+		}
+		return csvSchema;
+	}
+
+	@Override
+	public Optional<Encoder<RowData>> createEncoder(WriterContext context) {
+
+		CsvRowDataSerializationSchema.Builder builder = new CsvRowDataSerializationSchema.Builder(
+			context.getFormatRowType());
+
+		DescriptorProperties properties = getValidatedProperties(context.getFormatProperties());
+
+		properties.getOptionalCharacter(FORMAT_FIELD_DELIMITER)
+			.ifPresent(builder::setFieldDelimiter);
+
+		properties.getOptionalString(FORMAT_LINE_DELIMITER)
+			.ifPresent(builder::setLineDelimiter);
+
+		if (properties.getOptionalBoolean(FORMAT_DISABLE_QUOTE_CHARACTER).orElse(false)) {
+			builder.disableQuoteCharacter();
+		} else {
+			properties.getOptionalCharacter(FORMAT_QUOTE_CHARACTER).ifPresent(builder::setQuoteCharacter);
+		}
+
+		properties.getOptionalString(FORMAT_ARRAY_ELEMENT_DELIMITER)
+			.ifPresent(builder::setArrayElementDelimiter);
+
+		properties.getOptionalCharacter(FORMAT_ESCAPE_CHARACTER)
+			.ifPresent(builder::setEscapeCharacter);
+
+		properties.getOptionalString(FORMAT_NULL_LITERAL)
+			.ifPresent(builder::setNullLiteral);
+
+		return Optional.of(new CsvRowDataEncoder(builder.build()));
+	}
+
+	@Override
+	public Optional<BulkWriter.Factory<RowData>> createBulkWriterFactory(WriterContext context) {
+		return Optional.empty();
+	}
+
+	@Override
+	public boolean supportsSchemaDerivation() {
+		return true;
+	}
+
+	@Override
+	public List<String> supportedProperties() {
+		final List<String> properties = new ArrayList<>();
+		properties.add(FORMAT_FIELD_DELIMITER);
+		properties.add(FORMAT_LINE_DELIMITER);
+		properties.add(FORMAT_DISABLE_QUOTE_CHARACTER);
+		properties.add(FORMAT_QUOTE_CHARACTER);
+		properties.add(FORMAT_ALLOW_COMMENTS);
+		properties.add(FORMAT_IGNORE_PARSE_ERRORS);
+		properties.add(FORMAT_ARRAY_ELEMENT_DELIMITER);
+		properties.add(FORMAT_ESCAPE_CHARACTER);
+		properties.add(FORMAT_NULL_LITERAL);
+		return properties;
+	}
+
+	@Override
+	public Map<String, String> requiredContext() {
+		Map<String, String> context = new HashMap<>();
+		context.put(FORMAT, "csv");
+		return context;
+	}
+
+	private static DescriptorProperties getValidatedProperties(Map<String, String> propertiesMap) {
+		final DescriptorProperties properties = new DescriptorProperties(true);
+		properties.putProperties(propertiesMap);
+
+		properties.validateString(FORMAT_FIELD_DELIMITER, true, 1, 1);
+		properties.validateEnumValues(FORMAT_LINE_DELIMITER, true, Arrays.asList("\r", "\n", "\r\n", ""));
+		properties.validateBoolean(FORMAT_DISABLE_QUOTE_CHARACTER, true);
+		properties.validateString(FORMAT_QUOTE_CHARACTER, true, 1, 1);
+		properties.validateBoolean(FORMAT_ALLOW_COMMENTS, true);
+		properties.validateBoolean(FORMAT_IGNORE_PARSE_ERRORS, true);
+		properties.validateString(FORMAT_ARRAY_ELEMENT_DELIMITER, true, 1);
+		properties.validateString(FORMAT_ESCAPE_CHARACTER, true, 1, 1);
+
+		return properties;
+	}
+
+	/**
+	 * InputFormat that reads csv record into {@link RowData}.
+	 */
+	public static class CsvInputFormat extends AbstractCsvInputFormat<RowData> {
+		private static final long serialVersionUID = 1L;
+
+		private final RowType formatRowType;
+		private final DataType[] fieldTypes;
+		private final String[] fieldNames;
+		private final int[] selectFields;
+		private final List<String> partitionKeys;
+		private final String defaultPartValue;
+		private final long limit;
+		private final int[] csvSelectFieldToProjectFieldMapping;
+		private final int[] csvSelectFieldToCsvFieldMapping;
+		private final boolean ignoreParseErrors;
+
+		private transient InputStreamReader inputStreamReader;
+		private transient BufferedReader reader;
+		private transient boolean end;
+		private transient long emitted;
+		// reuse object for per record
+		private transient GenericRowData rowData;
+		private transient DeserializationRuntimeConverter runtimeConverter;
+		private transient MappingIterator<JsonNode> iterator;
+
+		public CsvInputFormat(
+				Path[] filePaths,
+				DataType[] fieldTypes,
+				String[] fieldNames,
+				CsvSchema csvSchema,
+				RowType formatRowType,
+				int[] selectFields,
+				List<String> partitionKeys,
+				String defaultPartValue,
+				long limit,
+				int[] csvSelectFieldToProjectFieldMapping,
+				int[] csvSelectFieldToCsvFieldMapping,
+				boolean ignoreParseErrors) {
+			super(filePaths, csvSchema);
+			this.fieldTypes = fieldTypes;
+			this.fieldNames = fieldNames;
+			this.formatRowType = formatRowType;
+			this.partitionKeys = partitionKeys;
+			this.defaultPartValue = defaultPartValue;
+			this.selectFields = selectFields;
+			this.limit = limit;
+			this.emitted = 0;
+			this.csvSelectFieldToProjectFieldMapping = csvSelectFieldToProjectFieldMapping;
+			this.csvSelectFieldToCsvFieldMapping = csvSelectFieldToCsvFieldMapping;
+			this.ignoreParseErrors = ignoreParseErrors;
+		}
+
+		@Override
+		public void open(FileInputSplit split) throws IOException {
+			super.open(split);
+			this.end = false;
+			this.inputStreamReader = new InputStreamReader(csvInputStream);
+			this.reader = new BufferedReader(inputStreamReader);
+			this.rowData = PartitionPathUtils.fillPartitionValueForRecord(fieldNames, fieldTypes, selectFields,
+				partitionKeys, currentSplit.getPath(), defaultPartValue);
+			this.iterator = new CsvMapper()
+				.readerFor(JsonNode.class)
+				.with(csvSchema)
+				.readValues(csvInputStream);
+			prepareRuntimeConverter();
+		}
+
+		private void prepareRuntimeConverter(){
+			CsvRowDataDeserializationSchema.Builder builder = new CsvRowDataDeserializationSchema.Builder(
+				formatRowType, new GenericTypeInfo<>(RowData.class))
+				.setIgnoreParseErrors(ignoreParseErrors);
+			this.runtimeConverter = builder.build().createRowConverter(formatRowType, true);
+		}
+
+		@Override
+		public boolean reachedEnd() throws IOException {
+			return emitted >= limit || end;
+		}
+
+		@Override
+		public RowData nextRecord(RowData reuse) throws IOException {
+			GenericRowData csvRow = null;
+			do {
+				try {
+					JsonNode root = iterator.nextValue();
+					csvRow = (GenericRowData) runtimeConverter.convert(root);
+				} catch (NoSuchElementException e) {
+					end = true;

Review comment:
       I update the logic, please review again.




----------------------------------------------------------------
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:
users@infra.apache.org