Skip to content

Conversation

@norberttech
Copy link
Member

Change Log

Added

Fixed

  • Parquet flattened for empty nested structures
  • Parquet RLEHybrid bitWidth calculation

Changed

  • Parquet - reading byte arrays now returns strings when option byte string as array is set

Removed

  • BytesToString data converted as it became redundant

Deprecated

Security


Description

\iterator_to_array((new Dremel())->assemble([1, 2], [1], [1, 2]));
}

public function test_decoding_nested_list_with_nulls_and_different_size_of_each_list() : void
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

values for those tests were generated with broken shredding logic, they are no longer valid and useful as it's pretty much impossible to understand whats going on here.

@github-actions
Copy link
Contributor

Flow PHP - Benchmarks

Results of the benchmarks from this PR are compared with the results from 1.x branch.

Extractors
+-----------------------+-------------------+------+-----+-------------------+-------------------+-----------------+
| benchmark             | subject           | revs | its | mem_peak          | mode              | rstdev          |
+-----------------------+-------------------+------+-----+-------------------+-------------------+-----------------+
| CSVExtractorBench     | bench_extract_10k | 1    | 3   | 3.914mb +0.03%    | 511.371ms +0.01%  | ±2.40% +12.73%  |
| JsonExtractorBench    | bench_extract_10k | 1    | 3   | 3.947mb +0.03%    | 1.072s -0.23%     | ±0.97% +6.66%   |
| ParquetExtractorBench | bench_extract_10k | 1    | 3   | 135.385mb -78.94% | 428.823ms -42.89% | ±0.71% +139.13% |
| TextExtractorBench    | bench_extract_10k | 1    | 3   | 3.674mb +0.03%    | 33.940ms -0.17%   | ±0.20% -56.42%  |
| XmlExtractorBench     | bench_extract_10k | 1    | 3   | 3.621mb +0.03%    | 437.850ms -0.02%  | ±0.83% -66.28%  |
+-----------------------+-------------------+------+-----+-------------------+-------------------+-----------------+
Transformers
+-----------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
| benchmark                   | subject                  | revs | its | mem_peak         | mode            | rstdev         |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
| RenameEntryTransformerBench | bench_transform_10k_rows | 1    | 3   | 115.966mb +0.00% | 60.742ms -0.00% | ±0.97% +50.54% |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
Loaders
+--------------------+----------------+------+-----+-------------------+-----------------+----------------+
| benchmark          | subject        | revs | its | mem_peak          | mode            | rstdev         |
+--------------------+----------------+------+-----+-------------------+-----------------+----------------+
| CSVLoaderBench     | bench_load_10k | 1    | 3   | 54.069mb +0.00%   | 85.567ms -0.10% | ±1.20% -17.47% |
| JsonLoaderBench    | bench_load_10k | 1    | 3   | 106.502mb +0.00%  | 53.170ms +0.46% | ±0.68% -32.18% |
| ParquetLoaderBench | bench_load_10k | 1    | 3   | 225.834mb -45.19% | 1.252s -17.14%  | ±0.38% -74.56% |
| TextLoaderBench    | bench_load_10k | 1    | 3   | 16.863mb +0.01%   | 44.680ms +0.65% | ±0.27% -17.11% |
+--------------------+----------------+------+-----+-------------------+-----------------+----------------+
Building Blocks
+-------------------------+----------------------------+------+-----+------------------+------------------+-----------------+
| benchmark               | subject                    | revs | its | mem_peak         | mode             | rstdev          |
+-------------------------+----------------------------+------+-----+------------------+------------------+-----------------+
| TypeDetectorBench       | bench_type_detector        | 1    | 3   | 59.695mb +0.00%  | 436.050ms -0.30% | ±0.83% -74.67%  |
| TypeDetectorBench       | bench_type_detector        | 1    | 3   | 14.234mb +0.01%  | 87.922ms +0.73%  | ±2.56% +265.96% |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 116.517mb +0.00% | 493.933ms -1.56% | ±0.32% -89.78%  |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 59.995mb +0.00%  | 251.460ms -0.60% | ±1.73% +40.78%  |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 14.928mb +0.01%  | 53.834ms -1.08%  | ±0.79% +708.66% |
| RowsBench               | bench_chunk_10_on_10k      | 2    | 3   | 86.786mb +0.00%  | 3.692ms +1.98%   | ±1.40% -38.44%  |
| RowsBench               | bench_diff_left_1k_on_10k  | 2    | 3   | 102.385mb +0.00% | 189.468ms -1.33% | ±0.95% -15.59%  |
| RowsBench               | bench_diff_right_1k_on_10k | 2    | 3   | 85.105mb +0.00%  | 19.457ms +1.01%  | ±0.54% -41.52%  |
| RowsBench               | bench_drop_1k_on_10k       | 2    | 3   | 88.026mb +0.00%  | 2.067ms -6.04%   | ±1.48% +230.29% |
| RowsBench               | bench_drop_right_1k_on_10k | 2    | 3   | 88.026mb +0.00%  | 2.010ms -9.51%   | ±2.69% +11.02%  |
| RowsBench               | bench_entries_on_10k       | 2    | 3   | 85.139mb +0.00%  | 3.166ms +7.23%   | ±2.64% -3.39%   |
| RowsBench               | bench_filter_on_10k        | 2    | 3   | 85.667mb +0.00%  | 16.843ms +9.27%  | ±0.49% -28.04%  |
| RowsBench               | bench_find_on_10k          | 2    | 3   | 85.667mb +0.00%  | 16.782ms +7.30%  | ±2.29% +61.54%  |
| RowsBench               | bench_find_one_on_10k      | 10   | 3   | 83.572mb +0.00%  | 2.006μs +0.61%   | ±2.32% -3.28%   |
| RowsBench               | bench_first_on_10k         | 10   | 3   | 83.572mb +0.00%  | 0.400μs +33.33%  | ±0.00% +0.00%   |
| RowsBench               | bench_flat_map_on_1k       | 2    | 3   | 92.922mb +0.00%  | 13.342ms +3.17%  | ±0.98% -67.04%  |
| RowsBench               | bench_map_on_10k           | 2    | 3   | 122.292mb +0.00% | 61.864ms -0.57%  | ±1.16% +21.00%  |
| RowsBench               | bench_merge_1k_on_10k      | 2    | 3   | 86.187mb +0.00%  | 1.670ms -8.06%   | ±3.51% +185.43% |
| RowsBench               | bench_partition_by_on_10k  | 2    | 3   | 89.534mb +0.00%  | 66.415ms +0.31%  | ±0.13% -86.17%  |
| RowsBench               | bench_remove_on_10k        | 2    | 3   | 88.289mb +0.00%  | 4.892ms +9.17%   | ±3.43% +39.02%  |
| RowsBench               | bench_sort_asc_on_1k       | 2    | 3   | 83.715mb +0.00%  | 42.548ms +4.25%  | ±0.12% -93.49%  |
| RowsBench               | bench_sort_by_on_1k        | 2    | 3   | 83.716mb +0.00%  | 41.696ms +2.59%  | ±1.00% -22.85%  |
| RowsBench               | bench_sort_desc_on_1k      | 2    | 3   | 83.715mb +0.00%  | 42.334ms +3.76%  | ±1.03% -59.67%  |
| RowsBench               | bench_sort_entries_on_1k   | 2    | 3   | 86.013mb +0.00%  | 7.667ms +1.35%   | ±1.76% +17.12%  |
| RowsBench               | bench_sort_on_1k           | 2    | 3   | 83.572mb +0.00%  | 30.243ms +1.09%  | ±1.40% -42.95%  |
| RowsBench               | bench_take_1k_on_10k       | 10   | 3   | 83.572mb +0.00%  | 14.225μs -9.98%  | ±3.43% +50.16%  |
| RowsBench               | bench_take_right_1k_on_10k | 10   | 3   | 83.572mb +0.00%  | 16.400μs -1.10%  | ±0.00% -100.00% |
| RowsBench               | bench_unique_on_1k         | 2    | 3   | 102.386mb +0.00% | 194.901ms +1.12% | ±0.81% +22.51%  |
+-------------------------+----------------------------+------+-----+------------------+------------------+-----------------+

@norberttech norberttech merged commit ab5af92 into flow-php:1.x Jul 15, 2024
@norberttech norberttech deleted the bug/parquet-empty-nested-structures branch December 5, 2024 20:19
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant