[GitHub] [flink] twalthr commented on a change in pull request #11692: [FLINK-16992][table-common] Add all ability interfaces for table sources and sinks

[GitBox <gi...@apache.org>]

twalthr commented on a change in pull request #11692: [FLINK-16992][table-common] Add all ability interfaces for table sources and sinks
URL: https://github.com/apache/flink/pull/11692#discussion_r408319531
 
 

 ##########
 File path: flink-table/flink-table-common/src/main/java/org/apache/flink/table/connector/source/abilities/SupportsProjectionPushDown.java
 ##########
 @@ -0,0 +1,72 @@
+/*
+ * 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.flink.table.connector.source.abilities;
+
+import org.apache.flink.annotation.PublicEvolving;
+import org.apache.flink.table.connector.source.ScanTableSource;
+
+/**
+ * Enables to push down a (possibly nested) projection into a {@link ScanTableSource}.
+ *
+ * <p>Given the following SQL:
+ * <pre>{@code
+ *   CREATE TABLE t (i INT, r ROW < d DOUBLE, b BOOLEAN>, s STRING);
+ *   SELECT s, r.d FROM t;
+ * }</pre>
+ *
+ * <p>In the above example, {@code r.d} and {@code s} are required fields. Other fields can be skipped
+ * in a projection.
+ *
+ * <p>By default, if this interface is not implemented, a projection is applied in a subsequent operation
+ * after the source.
+ *
+ * <p>For efficiency, a source can push a projection further down in order to be close to the actual
+ * data generation. A projection is only selecting fields that are used by a query. It does neither
+ * reorder fields nor contain any computation. A projection can either be performed on the fields of
+ * the top-level row only or consider nested fields as well (see {@link #supportsNestedProjection()}).
+ *
+ * <p>Note: If a source implements {@link SupportsComputedColumnPushDown}, the projection must be applied
+ * to the physically produced data only. {@link SupportsComputedColumnPushDown} assumes an already applied
+ * projection and includes a projection of computed columns.
+ */
+@PublicEvolving
+public interface SupportsProjectionPushDown {
+
+	/**
+	 * Returns whether this source supports nested projection.
+	 */
+	boolean supportsNestedProjection();
+
+	/**
+	 * Provides the field index paths that should be used for a projection. The indices are 0-based and
+	 * support fields within (possibly nested) structures if this is enabled via {@link #supportsNestedProjection()}.
+	 *
+	 * <p>In the example mentioned in {@link SupportsProjectionPushDown}, this method would receive:
+	 * <ul>
+	 *     <li>{@code [[1], [2]]} for {@code r} and {@code s} if {@link #supportsNestedProjection()}
+	 *     returns false.
+	 *     <li>{@code [[1, 0], [2]]} for {@code r.d} and {@code s} if {@link #supportsNestedProjection()}
+	 *     returns true.
+	 * </ul>
+	 *
+	 * @param projectedFields field index paths of all fields that must be present in the physically
+	 *                        produced data
+	 */
+	void applyProjection(int[][] projectedFields);
 
 Review comment:
   With this approach we can prune deeply nested fields and even reorder those fields. The argument is simply an array of paths for selection. If we prune `r.f.q1`, its path will not be in the array. For your example the array will look like:
   
   `[[0, 1, 1]]` which is equivalent to `[["r", "f", "q2"]]`. The latter can be simply computed out of a table schema or data type.

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services