bvaradar commented on a change in pull request #4910:
URL: https://github.com/apache/hudi/pull/4910#discussion_r836733080
##########
File path:
hudi-spark-datasource/hudi-spark-common/src/main/scala/org/apache/hudi/MergeOnReadSnapshotRelation.scala
##########
@@ -61,6 +61,11 @@ class MergeOnReadSnapshotRelation(sqlContext: SQLContext,
tableSchema: HoodieTableSchema,
requiredSchema: HoodieTableSchema,
filters: Array[Filter]):
HoodieMergeOnReadRDD = {
+ if (!internalSchema.isEmptySchema) {
+ // it is safe to enable vectorizedReader
+
sqlContext.sparkSession.sessionState.conf.setConfString("spark.sql.parquet.recordLevelFilter.enabled",
"false")
Review comment:
@xiarixiaoyao : Also, here
##########
File path:
hudi-common/src/main/java/org/apache/hudi/internal/schema/InternalSchema.java
##########
@@ -0,0 +1,291 @@
+/*
+ * 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.internal.schema;
+
+import org.apache.hudi.common.util.StringUtils;
+import org.apache.hudi.internal.schema.Types.Field;
+import org.apache.hudi.internal.schema.Types.RecordType;
+
+import java.io.Serializable;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.Comparator;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Locale;
+import java.util.Map;
+import java.util.Set;
+import java.util.stream.Collectors;
+
+/**
+ * Internal schema for hudi table.
+ * used to support schema evolution.
+ */
+public class InternalSchema implements Serializable {
+
+ private static final long DEFAULT_VERSION_ID = 0;
+
+ private final RecordType record;
+
+ private int maxColumnId;
+ private long versionId;
+
+ private transient Map<Integer, Field> idToField = null;
+ private transient Map<String, Integer> nameToId = null;
+ private transient Map<Integer, String> idToName = null;
+
+ public static InternalSchema getEmptyInternalSchema() {
+ return new InternalSchema(-1L, new ArrayList<>());
+ }
+
+ public boolean isEmptySchema() {
+ return versionId < 0;
+ }
+
+ public InternalSchema(List<Field> columns) {
+ this(DEFAULT_VERSION_ID, columns);
+ }
+
+ public InternalSchema(Field... columns) {
+ this(DEFAULT_VERSION_ID, Arrays.asList(columns));
+ }
+
+ public InternalSchema(long versionId, List<Field> cols) {
+ this.versionId = versionId;
+ this.record = RecordType.get(cols);
+ idToName = cols.isEmpty() ? new HashMap<>() :
InternalSchemaBuilder.getBuilder().buildIdToName(record);
+ nameToId = cols.isEmpty() ? new HashMap<>() :
idToName.entrySet().stream().collect(Collectors.toMap(Map.Entry::getValue,
Map.Entry::getKey));
+ maxColumnId = idToName.isEmpty() ? -1 :
idToName.keySet().stream().max(Comparator.comparing(Integer::valueOf)).get();
+ }
+
+ public InternalSchema(long versionId, int maxColumnId, List<Field> cols) {
+ this.maxColumnId = maxColumnId;
+ this.versionId = versionId;
+ this.record = RecordType.get(cols);
+ buildIdToName();
+ }
+
+ public InternalSchema(long versionId, int maxColumnId, Field... cols) {
+ this(versionId, maxColumnId, Arrays.asList(cols));
+ }
+
+ public RecordType getRecord() {
+ return record;
+ }
+
+ private Map<Integer, String> buildIdToName() {
+ if (idToName == null) {
+ idToName = InternalSchemaBuilder.getBuilder().buildIdToName(record);
+ }
+ return idToName;
+ }
+
+ private Map<String, Integer> buildNameToId() {
+ if (nameToId == null) {
+ if (idToName != null && !idToName.isEmpty()) {
+ nameToId =
idToName.entrySet().stream().collect(Collectors.toMap(Map.Entry::getValue,
Map.Entry::getKey));
+ return nameToId;
+ }
+ nameToId = InternalSchemaBuilder.getBuilder().buildNameToId(record);
+ }
+ return nameToId;
+ }
+
+ private Map<Integer, Field> buildIdToField() {
+ if (idToField == null) {
+ idToField = InternalSchemaBuilder.getBuilder().buildIdToField(record);
+ }
+ return idToField;
+ }
+
+ /**
+ * Get all columns full name.
+ */
+ public List<String> getAllColsFullName() {
+ if (nameToId == null) {
+ nameToId = InternalSchemaBuilder.getBuilder().buildNameToId(record);
+ }
+ return Arrays.asList(nameToId.keySet().toArray(new String[0]));
+ }
+
+ /**
+ * Set the version ID for this schema.
+ */
+ public InternalSchema setSchemaId(long versionId) {
+ this.versionId = versionId;
+ return this;
+ }
+
+ /**
+ * Returns the version ID for this schema.
+ */
+ public long schemaId() {
+ return this.versionId;
+ }
+
+ /**
+ * Set the version ID for this schema.
+ */
+ public void setMax_column_id(int maxColumnId) {
Review comment:
Use camel case for all names (class, methods, variables)
##########
File path:
hudi-spark-datasource/hudi-spark-common/src/main/scala/org/apache/hudi/BaseFileOnlyRelation.scala
##########
@@ -60,6 +60,12 @@ class BaseFileOnlyRelation(sqlContext: SQLContext,
tableSchema: HoodieTableSchema,
requiredSchema: HoodieTableSchema,
filters: Array[Filter]): HoodieUnsafeRDD =
{
+ if (!internalSchema.isEmptySchema) {
+ // it is safe to enable vectorizedReader
+
sqlContext.sparkSession.sessionState.conf.setConfString("spark.sql.parquet.recordLevelFilter.enabled",
"false")
Review comment:
@xiarixiaoyao : Can you add the comment in code why
recordLevelFilter.enabled is turned off in this case ?
--
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]
