ahmedabu98 commented on code in PR #40691: URL: https://github.com/apache/beam/pull/40691#discussion_r4362220032
########## sdks/java/io/iceberg/src/main/java/org/apache/beam/sdk/io/iceberg/PartitionPaths.java: ########## @@ -0,0 +1,90 @@ +/* + * 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.beam.sdk.io.iceberg; + +import java.util.Collection; +import java.util.HashSet; +import java.util.Set; +import java.util.TreeSet; +import org.apache.iceberg.PartitionField; +import org.apache.iceberg.PartitionSpec; +import org.checkerframework.checker.nullness.qual.Nullable; + +/** + * Partition values in a file's path. With prefix {@code gs://lake/orders/}, the file {@code + * gs://lake/orders/dt=2024-01-15/f.parquet} has {@code dt = 2024-01-15}. If the table is + * partitioned by {@code dt} as is (not by a transform such as {@code bucket}), readers take {@code + * dt} from that value, so the file needs no {@code dt} column. + */ +final class PartitionPaths { + /** + * Hive writes {@code dt=__HIVE_DEFAULT_PARTITION__} when dt is null; Iceberg reads it as null. + */ + static final String HIVE_NULL = "__HIVE_DEFAULT_PARTITION__"; + + private PartitionPaths() {} + + /** + * With prefix {@code gs://lake/orders/}: {@code gs://lake/orders/dt=1/region=EU/f.parquet} gives + * {@code dt=1/region=EU}, and {@code gs://lake/orders/f.parquet} gives "". + */ + static String relative(String filePath, String prefix) { + String partitionPath = filePath.substring(prefix.length()); + int lastSlashIndex = partitionPath.lastIndexOf('/'); + return lastSlashIndex > 0 ? partitionPath.substring(0, lastSlashIndex) : ""; + } + + /** + * The partition keys the file's folders give a value: {@code dt=1/region=EU/f.parquet} gives + * {@code [dt, region]}, {@code dt=__HIVE_DEFAULT_PARTITION__/region=EU/f.parquet} gives {@code + * [region]}, and {@code 2024/f.parquet} gives none. Also none without a prefix or for a file + * outside it. Needs no table, unlike Iceberg's parser, so the pre-pass can use it. + */ + static Set<String> partitionKeysInPath(String filePath, @Nullable String prefix) { + Set<String> keys = new TreeSet<>(); + if (prefix == null || prefix.isEmpty() || !filePath.startsWith(prefix)) { + return keys; + } + String relative = relative(filePath, prefix); + if (relative.isEmpty()) { + return keys; + } + for (String folder : relative.split("/", -1)) { + String[] parts = folder.split("=", 2); + if (parts.length == 2 && !parts[0].isEmpty() && !HIVE_NULL.equals(parts[1])) { + keys.add(parts[0]); + } + } + return keys; + } + + /** + * Ids of the table columns these keys fill. Only plain partitions count: a file under {@code + * id_bucket=3/} in a table partitioned by {@code bucket(id, 4)} is in bucket 3, which does not + * say what its {@code id} is. + */ + static Set<Integer> columnsFilledByKeys(PartitionSpec spec, Collection<String> keys) { Review Comment: nit: `identityColumnsFilledByKeys` ? ########## sdks/java/io/iceberg/src/main/java/org/apache/beam/sdk/io/iceberg/AddFiles.java: ########## @@ -811,21 +812,36 @@ private Verdict verify(String filePath, FileFormat format, @Nullable ParquetMeta } catch (Exception e) { return Verdict.error(UNREADABLE_SCHEMA_ERROR + errorMessage(e)); } - @Nullable String uncovered = uncoveredReason(fileSchema); + Set<Integer> partitionColumnsFromPath = partitionColumnsFromPath(filePath); Review Comment: i'd clarify naming of variables and methods here and other places that these are actually just the **identity** partition columns (e.g. `identityPartitionColumnsFromPath`) -- 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]
