aokolnychyi commented on code in PR #58272: URL: https://github.com/apache/spark/pull/58272#discussion_r3969732970
########## sql/catalyst/src/main/java/org/apache/spark/sql/connector/catalog/SupportsReportCatalogStatistics.java: ########## @@ -0,0 +1,55 @@ +/* + * 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.connector.catalog; + +import org.apache.spark.annotation.Evolving; +import org.apache.spark.sql.connector.read.Statistics; + +/** + * A mix-in interface for {@link Table}. Data sources can implement this interface to report + * table-level statistics to Spark from catalog metadata, without building a scan. + * <p> + * These statistics represent the full table and are not adjusted for any filter or projection + * pushdown. They are used by the optimizer during planning phases that occur before scan-level + * pushdown, such as join-type selection for LEFT SEMI and LEFT ANTI joins. Because no scan is + * built, the cost of computing these statistics must be low (for example, read from a snapshot + * summary map or table properties). + * <p> + * After scan pushdown is applied, SupportsReportStatistics on the resulting Scan is used for more + * accurate, filter-aware statistics. Connectors that support both interfaces provide a two-level + * view: catalog-level stats here, and refined scan-level stats after pushdown. + * <p> + * If any field is unavailable (e.g., no snapshot has been committed yet), the corresponding + * {@link java.util.OptionalLong} should be empty rather than a guess. + * + * @since 5.0.0 + */ +@Evolving +public interface SupportsReportCatalogStatistics extends Table { Review Comment: This seems like an escape hatch that would allow us to add more rules that access the stats before the planning, which would be really dangerous. -- 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]
