JingsongLi commented on code in PR #7313: URL: https://github.com/apache/paimon/pull/7313#discussion_r3442145918
########## paimon-spark/paimon-spark-common/src/main/java/org/apache/paimon/spark/procedure/CompactChainTableProcedure.java: ########## @@ -0,0 +1,213 @@ +/* + * 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.paimon.spark.procedure; + +import org.apache.paimon.CoreOptions; +import org.apache.paimon.partition.PartitionPredicate; +import org.apache.paimon.spark.commands.PaimonSparkWriter; +import org.apache.paimon.spark.util.ScanPlanHelper$; +import org.apache.paimon.spark.utils.SparkProcedureUtils; +import org.apache.paimon.table.ChainGroupReadTable; +import org.apache.paimon.table.FallbackReadFileStoreTable; +import org.apache.paimon.table.FileStoreTable; +import org.apache.paimon.table.source.Split; +import org.apache.paimon.utils.ParameterUtils; +import org.apache.paimon.utils.StringUtils; + +import org.apache.spark.sql.Dataset; +import org.apache.spark.sql.PaimonUtils; +import org.apache.spark.sql.Row; +import org.apache.spark.sql.catalyst.InternalRow; +import org.apache.spark.sql.connector.catalog.Identifier; +import org.apache.spark.sql.connector.catalog.TableCatalog; +import org.apache.spark.sql.execution.datasources.v2.DataSourceV2Relation; +import org.apache.spark.sql.functions; +import org.apache.spark.sql.types.Metadata; +import org.apache.spark.sql.types.StructField; +import org.apache.spark.sql.types.StructType; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import java.util.List; +import java.util.Map; + +import static org.apache.paimon.utils.Preconditions.checkArgument; +import static org.apache.spark.sql.types.DataTypes.BooleanType; +import static org.apache.spark.sql.types.DataTypes.StringType; + +/** + * Procedure to compact chain table. Usage: + * + * <pre><code> + * -- Compact chain table, overwrite default is false + * CALL sys.compact_chain_table(table => 'db.table', partition => 'dt="20250810",hour="22"', [overwrite => true]) + * </code></pre> + */ +public class CompactChainTableProcedure extends BaseProcedure { + + private static final Logger LOG = LoggerFactory.getLogger(CompactChainTableProcedure.class); + + private static final ProcedureParameter[] PARAMETERS = + new ProcedureParameter[] { + ProcedureParameter.required("table", StringType), + ProcedureParameter.required("partition", StringType), + ProcedureParameter.optional("overwrite", BooleanType) + }; + + private static final StructType OUTPUT_TYPE = + new StructType( + new StructField[] { + new StructField("result", BooleanType, false, Metadata.empty()) + }); + + protected CompactChainTableProcedure(TableCatalog tableCatalog) { + super(tableCatalog); + } + + @Override + public ProcedureParameter[] parameters() { + return PARAMETERS; + } + + @Override + public StructType outputType() { + return OUTPUT_TYPE; + } + + @Override + public InternalRow[] call(InternalRow args) { + Identifier tableIdent = toIdentifier(args.getString(0), PARAMETERS[0].name()); + String partitionStr = args.getString(1); + boolean overwrite = !args.isNullAt(2) && args.getBoolean(2); + checkArgument(StringUtils.isNotEmpty(partitionStr), "Partition string cannot be empty"); + + return modifyPaimonTable( + tableIdent, + t -> { + checkArgument( + new CoreOptions(t.options()).isChainTable(), + "compact_chain_table only supports chain table"); + checkArgument( + t instanceof FallbackReadFileStoreTable, + "Table %s is not a chain table", + tableIdent); + FallbackReadFileStoreTable table = (FallbackReadFileStoreTable) t; + checkArgument( + table.other() instanceof ChainGroupReadTable, + "Table %s is not a chain table", + tableIdent); + DataSourceV2Relation relation = createRelation(tableIdent); + boolean success = + execute( + (ChainGroupReadTable) table.other(), + relation, + partitionStr, + overwrite); + return new InternalRow[] {newInternalRow(success)}; + }); + } + + private boolean execute( + ChainGroupReadTable table, + DataSourceV2Relation relation, + String partitionStr, + boolean overwrite) { + String partition = SparkProcedureUtils.toWhere(partitionStr); + FileStoreTable snapshotTable = table.wrapped(); + + ChainGroupReadTable.ChainTableBatchScan scan = + (ChainGroupReadTable.ChainTableBatchScan) table.newScan(); + PartitionPredicate partitionPredicate = + SparkProcedureUtils.convertToPartitionPredicate( + partition, table.schema().logicalPartitionType(), spark(), relation); + + // Check if target partition already exists in snapshot branch + boolean partitionExists = checkPartitionExists(snapshotTable, partition, relation); + if (partitionExists) { + if (overwrite) { + scan.skipPreloadTargetSnapshot().withPartitionFilter(partitionPredicate); + LOG.info("Found existing partition {}, will overwrite it.", partition); + } else { + LOG.info( + "Partition {} already exists in snapshot branch, skipping compaction.", + partitionStr); + return false; + } + } else { + scan.withPartitionFilter(partitionPredicate); + } + + List<Split> splits = scan.plan().splits(); + if (splits.isEmpty()) { + LOG.warn( + "Table {} partition {} has no data to compact, skipping.", table, partitionStr); + return false; + } + + Dataset<Row> datasetForWrite = + PaimonUtils.createDataset( + spark(), + ScanPlanHelper$.MODULE$.createNewScanPlan( + splits.toArray(new Split[0]), relation)); + + PaimonSparkWriter writer = PaimonSparkWriter.apply(snapshotTable); + if (partitionExists) { + writer.writeBuilder().withOverwrite(); + } + Map<String, String> targetPartition = ParameterUtils.getPartitions(partitionStr).get(0); Review Comment: `SparkProcedureUtils.toWhere(partitionStr)` accepts semicolon-separated partitions, but the write path only uses `get(0)` and then rewrites every row to that first partition via `withColumn`. If a user passes multiple partitions, data from later partitions can be written into the first partition. Please either reject `partitionStr` values containing multiple partitions or handle each target partition separately. ########## paimon-spark/paimon-spark-common/src/main/java/org/apache/paimon/spark/procedure/CompactChainTableProcedure.java: ########## @@ -0,0 +1,213 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.paimon.spark.procedure; + +import org.apache.paimon.CoreOptions; +import org.apache.paimon.partition.PartitionPredicate; +import org.apache.paimon.spark.commands.PaimonSparkWriter; +import org.apache.paimon.spark.util.ScanPlanHelper$; +import org.apache.paimon.spark.utils.SparkProcedureUtils; +import org.apache.paimon.table.ChainGroupReadTable; +import org.apache.paimon.table.FallbackReadFileStoreTable; +import org.apache.paimon.table.FileStoreTable; +import org.apache.paimon.table.source.Split; +import org.apache.paimon.utils.ParameterUtils; +import org.apache.paimon.utils.StringUtils; + +import org.apache.spark.sql.Dataset; +import org.apache.spark.sql.PaimonUtils; +import org.apache.spark.sql.Row; +import org.apache.spark.sql.catalyst.InternalRow; +import org.apache.spark.sql.connector.catalog.Identifier; +import org.apache.spark.sql.connector.catalog.TableCatalog; +import org.apache.spark.sql.execution.datasources.v2.DataSourceV2Relation; +import org.apache.spark.sql.functions; +import org.apache.spark.sql.types.Metadata; +import org.apache.spark.sql.types.StructField; +import org.apache.spark.sql.types.StructType; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import java.util.List; +import java.util.Map; + +import static org.apache.paimon.utils.Preconditions.checkArgument; +import static org.apache.spark.sql.types.DataTypes.BooleanType; +import static org.apache.spark.sql.types.DataTypes.StringType; + +/** + * Procedure to compact chain table. Usage: + * + * <pre><code> + * -- Compact chain table, overwrite default is false + * CALL sys.compact_chain_table(table => 'db.table', partition => 'dt="20250810",hour="22"', [overwrite => true]) + * </code></pre> + */ +public class CompactChainTableProcedure extends BaseProcedure { + + private static final Logger LOG = LoggerFactory.getLogger(CompactChainTableProcedure.class); + + private static final ProcedureParameter[] PARAMETERS = + new ProcedureParameter[] { + ProcedureParameter.required("table", StringType), + ProcedureParameter.required("partition", StringType), + ProcedureParameter.optional("overwrite", BooleanType) + }; + + private static final StructType OUTPUT_TYPE = + new StructType( + new StructField[] { + new StructField("result", BooleanType, false, Metadata.empty()) + }); + + protected CompactChainTableProcedure(TableCatalog tableCatalog) { + super(tableCatalog); + } + + @Override + public ProcedureParameter[] parameters() { + return PARAMETERS; + } + + @Override + public StructType outputType() { + return OUTPUT_TYPE; + } + + @Override + public InternalRow[] call(InternalRow args) { + Identifier tableIdent = toIdentifier(args.getString(0), PARAMETERS[0].name()); + String partitionStr = args.getString(1); + boolean overwrite = !args.isNullAt(2) && args.getBoolean(2); + checkArgument(StringUtils.isNotEmpty(partitionStr), "Partition string cannot be empty"); + + return modifyPaimonTable( + tableIdent, + t -> { + checkArgument( + new CoreOptions(t.options()).isChainTable(), + "compact_chain_table only supports chain table"); + checkArgument( + t instanceof FallbackReadFileStoreTable, + "Table %s is not a chain table", + tableIdent); + FallbackReadFileStoreTable table = (FallbackReadFileStoreTable) t; + checkArgument( + table.other() instanceof ChainGroupReadTable, + "Table %s is not a chain table", + tableIdent); + DataSourceV2Relation relation = createRelation(tableIdent); + boolean success = + execute( + (ChainGroupReadTable) table.other(), + relation, + partitionStr, + overwrite); + return new InternalRow[] {newInternalRow(success)}; + }); + } + + private boolean execute( + ChainGroupReadTable table, + DataSourceV2Relation relation, + String partitionStr, + boolean overwrite) { + String partition = SparkProcedureUtils.toWhere(partitionStr); + FileStoreTable snapshotTable = table.wrapped(); + + ChainGroupReadTable.ChainTableBatchScan scan = + (ChainGroupReadTable.ChainTableBatchScan) table.newScan(); + PartitionPredicate partitionPredicate = + SparkProcedureUtils.convertToPartitionPredicate( + partition, table.schema().logicalPartitionType(), spark(), relation); + + // Check if target partition already exists in snapshot branch + boolean partitionExists = checkPartitionExists(snapshotTable, partition, relation); + if (partitionExists) { + if (overwrite) { + scan.skipPreloadTargetSnapshot().withPartitionFilter(partitionPredicate); + LOG.info("Found existing partition {}, will overwrite it.", partition); + } else { + LOG.info( + "Partition {} already exists in snapshot branch, skipping compaction.", + partitionStr); + return false; + } + } else { + scan.withPartitionFilter(partitionPredicate); + } + + List<Split> splits = scan.plan().splits(); + if (splits.isEmpty()) { + LOG.warn( + "Table {} partition {} has no data to compact, skipping.", table, partitionStr); + return false; + } + + Dataset<Row> datasetForWrite = + PaimonUtils.createDataset( + spark(), + ScanPlanHelper$.MODULE$.createNewScanPlan( + splits.toArray(new Split[0]), relation)); + + PaimonSparkWriter writer = PaimonSparkWriter.apply(snapshotTable); + if (partitionExists) { + writer.writeBuilder().withOverwrite(); Review Comment: This overwrite path relies on the table-level `dynamic-partition-overwrite` option because it calls `withOverwrite()` with an empty static partition. If the snapshot branch/table has `dynamic-partition-overwrite=false`, `FileStoreCommitImpl` will build a null overwrite filter and may overwrite the whole snapshot branch instead of only the requested `partitionStr`. Could we force dynamic partition overwrite here or pass the target partition map explicitly so the procedure's overwrite scope is independent of table configuration? -- 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]
