EnricoMi commented on a change in pull request #31905: URL: https://github.com/apache/spark/pull/31905#discussion_r655942526
########## File path: sql/core/src/main/scala/org/apache/spark/sql/Observation.scala ########## @@ -0,0 +1,170 @@ +/* + * 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.spark.sql + +import java.util.UUID +import java.util.concurrent.TimeUnit + +import org.apache.spark.sql.execution.QueryExecution +import org.apache.spark.sql.util.QueryExecutionListener + +/** + * Helper class to simplify usage of [[Dataset.observe(String, Column, Column*)]]: + * + * {{{ + * // Observe row count (rows) and highest id (maxid) in the Dataset while writing it + * val observation = Observation("my_metrics") + * val observed_ds = ds.observe(observation, count(lit(1)).as("rows"), max($"id").as("maxid")) + * observed_ds.write.parquet("ds.parquet") + * val metrics = observation.get + * }}} + * + * This collects the metrics while the first action is executed on the obseerved dataset. Subsequent + * actions do not modify the metrics returned by [[org.apache.spark.sql.Observation.get]]. Retrieval + * of the metric via [[org.apache.spark.sql.Observation.get]] blocks until the first action has + * finished and metrics become available. You can add a timeout to that blocking via + * [[org.apache.spark.sql.Observation.waitCompleted]]: + * + * {{{ + * if (observation.waitCompleted(100, TimeUnit.MILLISECONDS)) { + * observation.get + * } + * }}} + * + * This class does not support streaming datasets. + * + * @param name name of the metric + * @since 3.2.0 + */ +class Observation(name: String) { Review comment: This `Observation` is trying hiding multi-threading from the user for use cases where they simply want a metric after they have called an action on a Dataset. Adding async notion to the API does not contribute to that goal. The `waitCompleted` is an edge case, usually you simply call `Observation.get`. If the user wanted a Future, they could wrap the `Observation.get` with one. What you are proposing would mean that everyone else who does not want a `Future` would get a `Future` and has to call `.value` to get an `Option[Try[Row]]` and deal with this result where `Observation.get` now returns the `Row`. I do not see where the Future simplifies the API. Maybe you can sketch this out a bit more on how it would look like for the user of `Observation` so the benefits become clearer. -- 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: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
