|
| 1 | +// Licensed to the Apache Software Foundation (ASF) under one |
| 2 | +// or more contributor license agreements. See the NOTICE file |
| 3 | +// distributed with this work for additional information |
| 4 | +// regarding copyright ownership. The ASF licenses this file |
| 5 | +// to you under the Apache License, Version 2.0 (the |
| 6 | +// "License"); you may not use this file except in compliance |
| 7 | +// with the License. You may obtain a copy of the License at |
| 8 | +// |
| 9 | +// http://www.apache.org/licenses/LICENSE-2.0 |
| 10 | +// |
| 11 | +// Unless required by applicable law or agreed to in writing, |
| 12 | +// software distributed under the License is distributed on an |
| 13 | +// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY |
| 14 | +// KIND, either express or implied. See the License for the |
| 15 | +// specific language governing permissions and limitations |
| 16 | +// under the License. |
| 17 | + |
| 18 | +#include "arrow/io/memory.h" |
| 19 | +#include "arrow/testing/gtest_util.h" |
| 20 | +#include "parquet/arrow/reader.h" |
| 21 | +#include "parquet/file_reader.h" |
| 22 | +#include "parquet/file_rewriter.h" |
| 23 | +#ifdef _MSC_VER |
| 24 | +# pragma warning(push) |
| 25 | +// Disable forcing value to bool warnings |
| 26 | +# pragma warning(disable : 4800) |
| 27 | +#endif |
| 28 | + |
| 29 | +#include <memory> |
| 30 | + |
| 31 | +#include "gtest/gtest.h" |
| 32 | + |
| 33 | +#include "parquet/arrow/test_util.h" |
| 34 | +#include "parquet/platform.h" |
| 35 | +#include "parquet/properties.h" |
| 36 | + |
| 37 | +using arrow::Table; |
| 38 | +using arrow::io::BufferReader; |
| 39 | + |
| 40 | +namespace parquet::arrow { |
| 41 | + |
| 42 | +TEST(ParquetRewriterTest, ConcatRoundTrip) { |
| 43 | + auto rewriter_properties = |
| 44 | + RewriterProperties::Builder() |
| 45 | + .writer_properties( |
| 46 | + WriterProperties::Builder().enable_write_page_index()->build()) |
| 47 | + ->build(); |
| 48 | + |
| 49 | + auto schema = ::arrow::schema( |
| 50 | + {::arrow::field("a", ::arrow::int32()), ::arrow::field("b", ::arrow::utf8())}); |
| 51 | + |
| 52 | + std::shared_ptr<Buffer> buffer_up; |
| 53 | + std::shared_ptr<Buffer> buffer_down; |
| 54 | + |
| 55 | + WriteFile(rewriter_properties->writer_properties(), |
| 56 | + ::arrow::TableFromJSON(schema, {R"([[1, "a"], [2, "b"]])"}), buffer_up); |
| 57 | + WriteFile(rewriter_properties->writer_properties(), |
| 58 | + ::arrow::TableFromJSON(schema, {R"([[3, "c"]])"}), buffer_down); |
| 59 | + |
| 60 | + auto sink = CreateOutputStream(); |
| 61 | + auto rewriter = |
| 62 | + ParquetFileRewriter::Open({{std::make_shared<BufferReader>(buffer_up), |
| 63 | + std::make_shared<BufferReader>(buffer_down)}}, |
| 64 | + sink, {{NULLPTR, NULLPTR}}, NULLPTR, rewriter_properties); |
| 65 | + rewriter->Rewrite(); |
| 66 | + rewriter->Close(); |
| 67 | + |
| 68 | + ASSERT_OK_AND_ASSIGN(std::shared_ptr<Buffer> out_buffer, sink->Finish()); |
| 69 | + auto file_reader = ParquetFileReader::Open(std::make_shared<BufferReader>(out_buffer)); |
| 70 | + std::unique_ptr<FileReader> reader; |
| 71 | + ASSERT_OK_NO_THROW( |
| 72 | + FileReader::Make(::arrow::default_memory_pool(), std::move(file_reader), &reader)); |
| 73 | + |
| 74 | + std::shared_ptr<Table> table; |
| 75 | + ASSERT_OK(reader->ReadTable(&table)); |
| 76 | + ASSERT_OK(table->ValidateFull()); |
| 77 | + |
| 78 | + auto expected_table = |
| 79 | + ::arrow::TableFromJSON(schema, {R"([[1, "a"], [2, "b"], [3, "c"]])"}); |
| 80 | + AssertTablesEqual(*expected_table, *table); |
| 81 | +} |
| 82 | + |
| 83 | +TEST(ParquetRewriterTest, DISABLED_ExtendRoundTrip) { |
| 84 | + auto rewriter_properties = |
| 85 | + RewriterProperties::Builder() |
| 86 | + .writer_properties( |
| 87 | + WriterProperties::Builder().enable_write_page_index()->build()) |
| 88 | + ->build(); |
| 89 | + |
| 90 | + auto left_schema = ::arrow::schema( |
| 91 | + {::arrow::field("a", ::arrow::int32()), ::arrow::field("b", ::arrow::utf8())}); |
| 92 | + auto right_schema = ::arrow::schema({::arrow::field("c", ::arrow::int64())}); |
| 93 | + |
| 94 | + std::shared_ptr<Buffer> buffer_left; |
| 95 | + std::shared_ptr<Buffer> buffer_right; |
| 96 | + |
| 97 | + WriteFile(rewriter_properties->writer_properties(), |
| 98 | + ::arrow::TableFromJSON(left_schema, {R"([[1, "a"], [2, "b"], [3, "c"]])"}), |
| 99 | + buffer_left); |
| 100 | + WriteFile(rewriter_properties->writer_properties(), |
| 101 | + ::arrow::TableFromJSON(right_schema, {R"([[10], [20], [30]])"}), |
| 102 | + buffer_right); |
| 103 | + |
| 104 | + auto sink = CreateOutputStream(); |
| 105 | + auto rewriter = ParquetFileRewriter::Open( |
| 106 | + {{std::make_shared<BufferReader>(buffer_left)}, |
| 107 | + {std::make_shared<BufferReader>(buffer_right)}}, |
| 108 | + sink, {{NULLPTR}, {NULLPTR}}, NULLPTR, rewriter_properties); |
| 109 | + rewriter->Rewrite(); |
| 110 | + rewriter->Close(); |
| 111 | + |
| 112 | + ASSERT_OK_AND_ASSIGN(std::shared_ptr<Buffer> out_buffer, sink->Finish()); |
| 113 | + auto file_reader = ParquetFileReader::Open(std::make_shared<BufferReader>(out_buffer)); |
| 114 | + std::unique_ptr<FileReader> reader; |
| 115 | + ASSERT_OK_NO_THROW( |
| 116 | + FileReader::Make(::arrow::default_memory_pool(), std::move(file_reader), &reader)); |
| 117 | + |
| 118 | + std::shared_ptr<Table> table; |
| 119 | + ASSERT_OK(reader->ReadTable(&table)); |
| 120 | + ASSERT_OK(table->ValidateFull()); |
| 121 | + |
| 122 | + auto expected_schema = ::arrow::schema({::arrow::field("a", ::arrow::int32()), |
| 123 | + ::arrow::field("b", ::arrow::utf8()), |
| 124 | + ::arrow::field("c", ::arrow::int64())}); |
| 125 | + auto expected_table = ::arrow::TableFromJSON( |
| 126 | + expected_schema, {R"([[1, "a", 10], [2, "b", 20], [3, "c", 30]])"}); |
| 127 | + AssertTablesEqual(*expected_table, *table); |
| 128 | +} |
| 129 | + |
| 130 | +TEST(ParquetRewriterTest, DISABLED_SimpleRoundTrip) { |
| 131 | + auto rewriter_properties = RewriterProperties::Builder() |
| 132 | + .writer_properties(WriterProperties::Builder() |
| 133 | + .enable_write_page_index() |
| 134 | + ->max_row_group_length(1) |
| 135 | + ->build()) |
| 136 | + ->build(); |
| 137 | + |
| 138 | + auto left_schema = ::arrow::schema( |
| 139 | + {::arrow::field("a", ::arrow::int32()), ::arrow::field("b", ::arrow::utf8())}); |
| 140 | + auto right_schema = ::arrow::schema({::arrow::field("c", ::arrow::int64())}); |
| 141 | + |
| 142 | + std::shared_ptr<Buffer> buffer_left_up; |
| 143 | + std::shared_ptr<Buffer> buffer_left_down; |
| 144 | + std::shared_ptr<Buffer> buffer_right_up; |
| 145 | + std::shared_ptr<Buffer> buffer_right_down; |
| 146 | + |
| 147 | + WriteFile(rewriter_properties->writer_properties(), |
| 148 | + ::arrow::TableFromJSON(left_schema, {R"([[1, "a"], [2, "b"]])"}), |
| 149 | + buffer_left_up); |
| 150 | + WriteFile(rewriter_properties->writer_properties(), |
| 151 | + ::arrow::TableFromJSON(left_schema, {R"([[3, "c"]])"}), buffer_left_down); |
| 152 | + WriteFile(rewriter_properties->writer_properties(), |
| 153 | + ::arrow::TableFromJSON(right_schema, {R"([[10]])"}), buffer_right_up); |
| 154 | + WriteFile(rewriter_properties->writer_properties(), |
| 155 | + ::arrow::TableFromJSON(right_schema, {R"([[20], [30]])"}), buffer_right_down); |
| 156 | + |
| 157 | + auto sink = CreateOutputStream(); |
| 158 | + auto rewriter = ParquetFileRewriter::Open( |
| 159 | + {{std::make_shared<BufferReader>(buffer_left_up), |
| 160 | + std::make_shared<BufferReader>(buffer_left_down)}, |
| 161 | + {std::make_shared<BufferReader>(buffer_right_up), |
| 162 | + std::make_shared<BufferReader>(buffer_right_down)}}, |
| 163 | + sink, {{NULLPTR, NULLPTR}, {NULLPTR, NULLPTR}}, NULLPTR, rewriter_properties); |
| 164 | + rewriter->Rewrite(); |
| 165 | + rewriter->Close(); |
| 166 | + |
| 167 | + ASSERT_OK_AND_ASSIGN(std::shared_ptr<Buffer> out_buffer, sink->Finish()); |
| 168 | + auto file_reader = ParquetFileReader::Open(std::make_shared<BufferReader>(out_buffer)); |
| 169 | + std::unique_ptr<FileReader> reader; |
| 170 | + ASSERT_OK_NO_THROW( |
| 171 | + FileReader::Make(::arrow::default_memory_pool(), std::move(file_reader), &reader)); |
| 172 | + |
| 173 | + std::shared_ptr<Table> table; |
| 174 | + ASSERT_OK(reader->ReadTable(&table)); |
| 175 | + ASSERT_OK(table->ValidateFull()); |
| 176 | + |
| 177 | + auto expected_schema = ::arrow::schema({::arrow::field("a", ::arrow::int32()), |
| 178 | + ::arrow::field("b", ::arrow::utf8()), |
| 179 | + ::arrow::field("c", ::arrow::int64())}); |
| 180 | + auto expected_table = ::arrow::TableFromJSON( |
| 181 | + expected_schema, {R"([[1, "a", 10], [2, "b", 20], [3, "c", 30]])"}); |
| 182 | + AssertTablesEqual(*expected_table, *table); |
| 183 | +} |
| 184 | + |
| 185 | +} // namespace parquet::arrow |
0 commit comments