aokolnychyi commented on code in PR #6012: URL: https://github.com/apache/iceberg/pull/6012#discussion_r1034319524
########## spark/v3.3/spark/src/main/java/org/apache/iceberg/spark/procedures/GenerateChangesProcedure.java: ########## @@ -0,0 +1,336 @@ +/* + * 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.iceberg.spark.procedures; + +import static org.apache.iceberg.ChangelogOperation.DELETE; +import static org.apache.iceberg.ChangelogOperation.INSERT; + +import java.util.Arrays; +import java.util.Iterator; +import java.util.List; +import java.util.Objects; +import java.util.stream.Collectors; +import org.apache.iceberg.ChangelogOperation; +import org.apache.iceberg.MetadataColumns; +import org.apache.iceberg.Snapshot; +import org.apache.iceberg.Table; +import org.apache.iceberg.relocated.com.google.common.collect.Iterators; +import org.apache.iceberg.spark.SparkReadOptions; +import org.apache.iceberg.spark.source.SparkChangelogTable; +import org.apache.iceberg.util.DateTimeUtil; +import org.apache.iceberg.util.SnapshotUtil; +import org.apache.spark.api.java.function.MapPartitionsFunction; +import org.apache.spark.sql.Column; +import org.apache.spark.sql.DataFrameReader; +import org.apache.spark.sql.Dataset; +import org.apache.spark.sql.Row; +import org.apache.spark.sql.RowFactory; +import org.apache.spark.sql.catalyst.InternalRow; +import org.apache.spark.sql.catalyst.encoders.RowEncoder; +import org.apache.spark.sql.catalyst.expressions.GenericInternalRow; +import org.apache.spark.sql.catalyst.expressions.GenericRowWithSchema; +import org.apache.spark.sql.connector.catalog.Identifier; +import org.apache.spark.sql.connector.catalog.TableCatalog; +import org.apache.spark.sql.connector.iceberg.catalog.ProcedureParameter; +import org.apache.spark.sql.types.DataTypes; +import org.apache.spark.sql.types.Metadata; +import org.apache.spark.sql.types.StructField; +import org.apache.spark.sql.types.StructType; +import org.apache.spark.unsafe.types.UTF8String; +import org.jetbrains.annotations.NotNull; + +public class GenerateChangesProcedure extends BaseProcedure { + + public static SparkProcedures.ProcedureBuilder builder() { + return new BaseProcedure.Builder<GenerateChangesProcedure>() { + @Override + protected GenerateChangesProcedure doBuild() { + return new GenerateChangesProcedure(tableCatalog()); + } + }; + } + + private static final ProcedureParameter[] PARAMETERS = + new ProcedureParameter[] { + ProcedureParameter.required("table", DataTypes.StringType), + // the snapshot ids input are ignored when the start/end timestamps are provided + ProcedureParameter.optional("start_snapshot_id_exclusive", DataTypes.LongType), + ProcedureParameter.optional("end_snapshot_id_inclusive", DataTypes.LongType), + ProcedureParameter.optional("table_change_view", DataTypes.StringType), + ProcedureParameter.optional("identifier_columns", DataTypes.StringType), + ProcedureParameter.optional("start_timestamp", DataTypes.TimestampType), + ProcedureParameter.optional("end_timestamp", DataTypes.TimestampType), + }; + + private static final StructType OUTPUT_TYPE = + new StructType( + new StructField[] { + new StructField("view_name", DataTypes.StringType, false, Metadata.empty()) + }); + + private GenerateChangesProcedure(TableCatalog tableCatalog) { + super(tableCatalog); + } + + @Override + public ProcedureParameter[] parameters() { + return PARAMETERS; + } + + @Override + public StructType outputType() { + return OUTPUT_TYPE; + } + + @Override + public InternalRow[] call(InternalRow args) { + String tableName = args.getString(0); + + // Read data from the table.changes + Dataset<Row> df = changelogRecords(tableName, args); + + // Compute the pre-image and post-images if the identifier columns are provided. + if (!args.isNullAt(4)) { + String[] identifierColumns = args.getString(4).split(","); + if (identifierColumns.length > 0) { + df = withUpdate(df, identifierColumns); + } + } + + String viewName = viewName(args, tableName); + + // Create a view for users to query + df.createOrReplaceTempView(viewName); + + return toOutputRows(viewName); + } + + private Dataset<Row> changelogRecords(String tableName, InternalRow args) { + Long[] snapshotIds = getSnapshotIds(tableName, args); + + // we don't have to validate the snapshot ids here because the reader will do it for us. + DataFrameReader reader = spark().read(); + if (snapshotIds[0] != null) { + reader = reader.option(SparkReadOptions.START_SNAPSHOT_ID, snapshotIds[0]); + } + + if (snapshotIds[1] != null) { + reader = reader.option(SparkReadOptions.END_SNAPSHOT_ID, snapshotIds[1]); + } + + return reader.table(tableName + "." + SparkChangelogTable.TABLE_NAME); + } + + @NotNull + private Long[] getSnapshotIds(String tableName, InternalRow args) { + Long[] snapshotIds = new Long[] {null, null}; + + Long startTimestamp = args.isNullAt(5) ? null : DateTimeUtil.microsToMillis(args.getLong(5)); + Long endTimestamp = args.isNullAt(6) ? null : DateTimeUtil.microsToMillis(args.getLong(6)); + + if (startTimestamp == null && endTimestamp == null) { + snapshotIds[0] = args.isNullAt(1) ? null : args.getLong(1); + snapshotIds[1] = args.isNullAt(2) ? null : args.getLong(2); + } else { + Identifier tableIdent = toIdentifier(tableName, PARAMETERS[0].name()); + Table table = loadSparkTable(tableIdent).table(); + Snapshot[] snapshots = snapshotsFromTimestamp(startTimestamp, endTimestamp, table); + + if (snapshots != null) { + snapshotIds[0] = snapshots[0].parentId(); + snapshotIds[1] = snapshots[1].snapshotId(); + } + } + return snapshotIds; + } + + private Snapshot[] snapshotsFromTimestamp(Long startTimestamp, Long endTimestamp, Table table) { + Snapshot[] snapshots = new Snapshot[] {null, null}; + + if (startTimestamp != null && endTimestamp != null && startTimestamp > endTimestamp) { + throw new IllegalArgumentException( + "Start timestamp must be less than or equal to end timestamp"); + } + + if (startTimestamp == null) { + snapshots[0] = SnapshotUtil.oldestAncestor(table); + } else { + snapshots[0] = SnapshotUtil.oldestAncestorAfter(table, startTimestamp); + } + + if (endTimestamp == null) { + snapshots[1] = table.currentSnapshot(); + } else { + snapshots[1] = table.snapshot(SnapshotUtil.snapshotIdAsOfTime(table, endTimestamp)); + } + + if (snapshots[0] == null || snapshots[1] == null) { + return null; + } + + return snapshots; + } + + @NotNull + private static String viewName(InternalRow args, String tableName) { + String viewName = args.isNullAt(3) ? null : args.getString(3); + if (viewName == null) { + String shortTableName = + tableName.contains(".") ? tableName.substring(tableName.lastIndexOf(".") + 1) : tableName; + viewName = shortTableName + "_changes"; + } + return viewName; + } + + private Dataset<Row> withUpdate(Dataset<Row> df, String[] identifiers) { + Column[] partitionSpec = getPartitionSpec(df, identifiers); + Column[] sortSpec = sortSpec(df, partitionSpec); + + int changeTypeIdx = df.schema().fieldIndex(MetadataColumns.CHANGE_TYPE.name()); + List<Integer> partitionIdx = + Arrays.stream(partitionSpec) + .map(column -> df.schema().fieldIndex(column.toString())) + .collect(Collectors.toList()); + + return df.repartition(partitionSpec) + .sortWithinPartitions(sortSpec) + .mapPartitions( + processRowsWithinTask(changeTypeIdx, partitionIdx), RowEncoder.apply(df.schema())); + } + + @NotNull + private static Column[] getPartitionSpec(Dataset<Row> df, String[] identifiers) { + Column[] partitionSpec = new Column[identifiers.length + 1]; + for (int i = 0; i < identifiers.length; i++) { + try { + partitionSpec[i] = df.col(identifiers[i]); + } catch (Exception e) { + throw new IllegalArgumentException( + String.format("Identifier column '%s' does not exist in the table", identifiers[i]), e); + } + } + partitionSpec[partitionSpec.length - 1] = df.col(MetadataColumns.CHANGE_ORDINAL.name()); + return partitionSpec; + } + + @NotNull + private static Column[] sortSpec(Dataset<Row> df, Column[] partitionSpec) { + Column[] sortSpec = new Column[partitionSpec.length + 1]; + System.arraycopy(partitionSpec, 0, sortSpec, 0, partitionSpec.length); + sortSpec[sortSpec.length - 1] = df.col(MetadataColumns.CHANGE_TYPE.name()); + return sortSpec; + } + + private static MapPartitionsFunction<Row, Row> processRowsWithinTask( + int changeTypeIndex, List<Integer> partitionIdx) { + return rowIterator -> { + Iterator<Row> iterator = + new Iterator<Row>() { + private Row nextRow = null; + + @Override + public boolean hasNext() { + if (nextRow != null) { + return true; + } + return rowIterator.hasNext(); + } + + @Override + public Row next() { + // if the next row is cached and changed, return it directly + if (nextRow != null + && !nextRow.getString(changeTypeIndex).equals(DELETE.name()) + && !nextRow.getString(changeTypeIndex).equals(INSERT.name())) { + Row row = nextRow; + nextRow = null; + return row; + } + + Row currentRow = currentRow(); + + if (rowIterator.hasNext()) { + GenericRowWithSchema nextRow = (GenericRowWithSchema) rowIterator.next(); + + if (withinPartition(currentRow, nextRow) + && currentRow.getString(changeTypeIndex).equals(DELETE.name()) + && nextRow.getString(changeTypeIndex).equals(INSERT.name())) { Review Comment: I think creating a separate class for this functionality will allow us to unit test it. -- 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: issues-unsubscr...@iceberg.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org For additional commands, e-mail: issues-h...@iceberg.apache.org
