bryanck commented on code in PR #8555: URL: https://github.com/apache/iceberg/pull/8555#discussion_r1333250759
########## flink/v1.17/flink/src/main/java/org/apache/iceberg/flink/sink/CachingTableSupplier.java: ########## @@ -0,0 +1,84 @@ +/* + * 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.flink.sink; + +import java.time.Duration; +import org.apache.flink.util.Preconditions; +import org.apache.iceberg.Table; +import org.apache.iceberg.flink.TableLoader; +import org.apache.iceberg.util.DateTimeUtil; +import org.apache.iceberg.util.SerializableSupplier; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +/** + * A table loader that will only reload a table after a certain interval has passed. WARNING: This + * table loader should be used carefully when used with writer tasks. It could result in heavy load + * on a catalog for jobs with many writers. + */ +class CachingTableSupplier implements SerializableSupplier<Table> { + + private static final Logger LOG = LoggerFactory.getLogger(CachingTableSupplier.class); + + private final Table initialTable; + private final TableLoader tableLoader; + private final Duration tableRefreshInterval; + private long nextReloadTimeMs; + private transient Table table; + + CachingTableSupplier(Table initialTable, TableLoader tableLoader, Duration tableRefreshInterval) { + Preconditions.checkArgument(initialTable != null, "initialTable cannot be null"); + Preconditions.checkArgument(tableLoader != null, "tableLoader cannot be null"); + Preconditions.checkArgument( + tableRefreshInterval != null, "tableRefreshInterval cannot be null"); + this.initialTable = initialTable; + this.table = initialTable; + this.tableLoader = tableLoader; + this.tableRefreshInterval = tableRefreshInterval; + this.nextReloadTimeMs = System.currentTimeMillis() + tableRefreshInterval.toMillis(); + } + + @Override + public Table get() { + if (table == null) { + this.table = initialTable; + } + return table; + } + + public void refresh() { Review Comment: My main concern before was within a single subtask, rather than across different subtasks. I felt changing the schema or spec while the file is being written could potentially lead to problems down the road, where the refresh isn't done in a deterministic way. Even using a broadcast or token delegation framework, you could still cache and refresh in the supplier only at given times in the supplier for those. And before the latest change, no public APIs were changed. Either way, implementing schema evolution is going to require additional changes, so the current PR will be fine to refresh the FileIO and how the refresh occurs can be updated as needed when schema evolution is added. -- 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]
