amogh-jahagirdar commented on code in PR #12450: URL: https://github.com/apache/iceberg/pull/12450#discussion_r2093745842
########## api/src/main/java/org/apache/iceberg/actions/ComputePartitionStats.java: ########## @@ -0,0 +1,43 @@ +/* + * 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.actions; + +import org.apache.iceberg.PartitionStatisticsFile; + +/** + * An action that computes and writes the partition statistics of an Iceberg table. Current snapshot + * is used by default. + */ +public interface ComputePartitionStats + extends Action<ComputePartitionStats, ComputePartitionStats.Result> { Review Comment: I'm late to this PR so my apologies if this has already been discussed in the past but have we considered just including the ability for computing partition stats to the existing `ComputeTableStats` Action? I'm a bit wary of adding too many actions, especially in this case where an external user who is using spark has to know that there are 2 separate actions for "stats". In my head something like "compute table stats" + a partition stats option API on that seems better. I know the actual stats files are of course different but I'm mainly thinking from an action and API perspective, not exposing too many different concepts. Of course if we do try to add flags to the existing procedure we have to think through what's the default behavior when not specified (preserving compatibility) and then what additional modes we have. Thoughts @ajantha-bhat @karuppayya @nastra @aokolnychyi @rdblue ? ########## core/src/main/java/org/apache/iceberg/PartitionStatsHandler.java: ########## @@ -165,6 +165,14 @@ public static PartitionStatisticsFile computeAndWriteStatsFile(Table table, long "Using full compute as previous statistics file is not present for incremental compute."); stats = computeStats(table, snapshot, file -> true, false /* incremental */).values(); } else { + if (statisticsFile.snapshotId() == snapshotId) { + // no-op + LOG.info( + "Snapshot {} already has statistics file computed. Returning the same file as result.", Review Comment: "Using existing statistics file for snapshot {}" probably suffices. ########## spark/v3.5/spark/src/main/java/org/apache/iceberg/spark/actions/ComputePartitionStatsSparkAction.java: ########## @@ -0,0 +1,102 @@ +/* + * 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.actions; + +import java.io.IOException; +import org.apache.iceberg.PartitionStatisticsFile; +import org.apache.iceberg.PartitionStatsHandler; +import org.apache.iceberg.Snapshot; +import org.apache.iceberg.Table; +import org.apache.iceberg.actions.ComputePartitionStats; +import org.apache.iceberg.actions.ImmutableComputePartitionStats; +import org.apache.iceberg.relocated.com.google.common.base.Preconditions; +import org.apache.iceberg.spark.JobGroupInfo; +import org.apache.spark.sql.SparkSession; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +/** + * Computes the stats incrementally after the snapshot that has partition stats file till the given + * snapshot (uses current snapshot if not specified) and writes the combined result into a {@link + * PartitionStatisticsFile} after merging the stats for a given snapshot. Does a full compute if + * previous statistics file does not exist. Also registers the {@link PartitionStatisticsFile} to + * table metadata. + */ +public class ComputePartitionStatsSparkAction + extends BaseSparkAction<ComputePartitionStatsSparkAction> implements ComputePartitionStats { + + private static final Logger LOG = LoggerFactory.getLogger(ComputePartitionStatsSparkAction.class); + private static final Result EMPTY_RESULT = + ImmutableComputePartitionStats.Result.builder().build(); + + private final Table table; + private Snapshot snapshot; + + ComputePartitionStatsSparkAction(SparkSession spark, Table table) { + super(spark); + this.table = table; + this.snapshot = table.currentSnapshot(); + } + + @Override + protected ComputePartitionStatsSparkAction self() { + return this; + } + + @Override + public ComputePartitionStats snapshot(long newSnapshotId) { + Snapshot newSnapshot = table.snapshot(newSnapshotId); + Preconditions.checkArgument(newSnapshot != null, "Snapshot not found: %s", newSnapshotId); + this.snapshot = newSnapshot; + return this; + } + + @Override + public Result execute() { + if (snapshot == null) { + LOG.info("No snapshot to compute partition stats for table {}", table.name()); + return EMPTY_RESULT; + } + + JobGroupInfo info = newJobGroupInfo("COMPUTE-PARTITION-STATS", jobDesc()); + return withJobGroupInfo(info, this::doExecute); + } + + private Result doExecute() { + LOG.info("Computing partition stats for {} (snapshot {})", table.name(), snapshot.snapshotId()); Review Comment: I think I'd prefer without = and we can probably get away from the nested parentheses in a Log statement. "Computing partition stats for table {} snapshot {}" is sufficient info imo -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
