nastra commented on code in PR #13933: URL: https://github.com/apache/iceberg/pull/13933#discussion_r2413743360
########## api/src/main/java/org/apache/iceberg/stats/StatsUtil.java: ########## @@ -0,0 +1,223 @@ +/* + * 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.stats; + +import static org.apache.iceberg.types.Types.NestedField.optional; + +import java.util.Comparator; +import java.util.List; +import java.util.Objects; +import java.util.Set; +import java.util.stream.Collectors; +import org.apache.iceberg.Schema; +import org.apache.iceberg.relocated.com.google.common.collect.Lists; +import org.apache.iceberg.relocated.com.google.common.collect.Sets; +import org.apache.iceberg.types.Type; +import org.apache.iceberg.types.TypeUtil; +import org.apache.iceberg.types.Types; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +public class StatsUtil { + private static final Logger LOG = LoggerFactory.getLogger(StatsUtil.class); + static final int VALUE_COUNT_OFFSET = 0; + static final int NULL_VALUE_COUNT_OFFSET = 1; + static final int NAN_VALUE_COUNT_OFFSET = 2; + static final int AVG_VALUE_SIZE_OFFSET = 3; + static final int MAX_VALUE_SIZE_OFFSET = 4; + static final int LOWER_BOUND_OFFSET = 5; + static final int UPPER_BOUND_OFFSET = 6; + static final int NUM_STATS_PER_COLUMN = 200; + static final int RESERVED_FIELD_IDS = 200; + static final int DATA_SPACE_FIELD_ID_START = 10_000; + static final int METADATA_SPACE_FIELD_ID_START = 2_147_000_000; + static final int RESERVED_FIELD_IDS_START = Integer.MAX_VALUE - RESERVED_FIELD_IDS; + // support stats for only up to this amount of data field IDs + static final int MAX_DATA_FIELD_ID = 1_000_000; + static final int MAX_DATA_STATS_FIELD_ID = 200_010_000; + + private StatsUtil() {} + + public static int statsFieldIdForField(int fieldId) { + return fieldId >= RESERVED_FIELD_IDS_START + ? statsFieldIdForReservedField(fieldId) + : statsFieldIdForDataField(fieldId); + } + + private static int statsFieldIdForDataField(int fieldId) { + long statsFieldId = DATA_SPACE_FIELD_ID_START + NUM_STATS_PER_COLUMN * (long) fieldId; + if (fieldId < 0 || fieldId > MAX_DATA_FIELD_ID || statsFieldId < 0) { + return -1; + } + + return (int) statsFieldId; + } + + private static int statsFieldIdForReservedField(int fieldId) { + int offset = RESERVED_FIELD_IDS - (Integer.MAX_VALUE - fieldId); + + long statsFieldId = METADATA_SPACE_FIELD_ID_START + NUM_STATS_PER_COLUMN * (long) offset; + if (statsFieldId < 0 || statsFieldId > RESERVED_FIELD_IDS_START) { + // ID overflows + return -1; + } + + return (int) statsFieldId; + } + + public static int fieldIdForStatsField(int statsFieldId) { + if (statsFieldId < DATA_SPACE_FIELD_ID_START || statsFieldId % NUM_STATS_PER_COLUMN != 0) { + return -1; Review Comment: purely from the Java impl we should never hit this condition, because we control how stats field IDs are generated. This is mostly to prevent us from cases where e.g. other Iceberg clients accidentally used wrong stats field IDs. If that ever happens and we throw here, the manifest will be unreadable, which is what we want to avoid. Thus we need to be lenient on enforcing such things. We took a similar approach when implementing views, where we are lenient when reading potentially invalid metadata that doesn't fully adhere to the Spec just to avoid having unreadable metadata. That being said, we should do the same here, does that make sense? -- 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]
