wgtmac commented on code in PR #45234: URL: https://github.com/apache/arrow/pull/45234#discussion_r1947413710
########## cpp/src/parquet/row_range.h: ########## @@ -0,0 +1,87 @@ +// 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. + +#pragma once + +#include <variant> +#include <vector> + +#include "parquet/platform.h" + +namespace parquet { + +/// RowRanges is a collection of non-overlapping and ascendingly ordered row ranges. +class PARQUET_EXPORT RowRanges { + public: + /// \brief A range of contiguous rows represented by an interval. + struct IntervalRange { + /// Start row of the range (inclusive). + int64_t start; + /// End row of the range (inclusive). + int64_t end; + }; + + /// \brief A range of contiguous rows represented by a bitmap. + struct BitmapRange { Review Comment: > 1. at which use case is this API targeted? a. Specify the required row ranges when creating a Parquet reader (be it a `ParquetFileReader` or `RowGroupReader`). This applies to both interval and bitmap forms depending on the use case. Perhaps this is the main reason to use a mixed form. b. Used internally when computing matched row ranges from filter pushdown evaluation result with the help of page index. This applies to the interval range because operations are usually based on the page level. c. Advanced row-level filtering discussed in https://github.com/apache/arrow/issues/37559. This should be bitmap range used by `ColumnReader` and `Decoder`. > 2. what is the typical bitmap length this is meant to represent? It depends on the use case above and can be any of `num_rows_in_file`, `num_rows_in_a_row_group` and `batch_size`. > 3. do we want a structure to select across an entire row group, or is it better to have one selection per data page, or perhaps a two-level selection (with a page bitmap and then a per-page bitmap for those pages that haven't been pruned)? I don't think we need an intra-page selection since page boundaries are usually not aligned across different columns. I think `file-level selection` is easy to use for users and `row-group-level selection` is required by the `RowGroupReader` implementation. -- 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]
