Dremel & Parquet performance improvements

[norberttech]

Change Log

Added

Fixed

• Dremel & Parquet performance improvements

Changed

Removed

Deprecated

Security

---

Description

Profiling:

Before
After

[github-actions]

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          |
+-----------------------+-------------------+------+-----+------------------+------------------+-----------------+
| AvroExtractorBench    | bench_extract_10k | 1    | 3   | 35.123mb +0.00%  | 711.168ms +1.29% | ±2.41% +62.75%  |
| CSVExtractorBench     | bench_extract_10k | 1    | 3   | 4.822mb +0.01%   | 302.082ms +0.13% | ±0.32% +170.83% |
| JsonExtractorBench    | bench_extract_10k | 1    | 3   | 4.920mb +0.01%   | 934.940ms +2.06% | ±0.32% -19.76%  |
| ParquetExtractorBench | bench_extract_10k | 1    | 3   | 239.626mb +0.00% | 1.126s +1.63%    | ±0.28% -57.39%  |
| TextExtractorBench    | bench_extract_10k | 1    | 3   | 4.699mb +0.01%   | 24.820ms +0.40%  | ±0.07% -95.22%  |
| XmlExtractorBench     | bench_extract_10k | 1    | 3   | 4.700mb +0.01%   | 407.981ms -0.27% | ±0.14% -92.85%  |
+-----------------------+-------------------+------+-----+------------------+------------------+-----------------+

Transformers

+-----------------------------+--------------------------+------+-----+------------------+-----------------+-----------------+
| benchmark                   | subject                  | revs | its | mem_peak         | mode            | rstdev          |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+-----------------+
| RenameEntryTransformerBench | bench_transform_10k_rows | 1    | 3   | 110.386mb +0.00% | 65.139ms +2.29% | ±2.11% +137.99% |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+-----------------+

Loaders

+--------------------+----------------+------+-----+------------------+------------------+-----------------+
| benchmark          | subject        | revs | its | mem_peak         | mode             | rstdev          |
+--------------------+----------------+------+-----+------------------+------------------+-----------------+
| AvroLoaderBench    | bench_load_10k | 1    | 3   | 94.784mb +0.00%  | 444.697ms -2.91% | ±1.98% +54.59%  |
| CSVLoaderBench     | bench_load_10k | 1    | 3   | 54.862mb +0.00%  | 71.688ms -0.12%  | ±0.27% -53.13%  |
| JsonLoaderBench    | bench_load_10k | 1    | 3   | 105.445mb +0.00% | 58.582ms +3.18%  | ±1.29% +315.07% |
| ParquetLoaderBench | bench_load_10k | 1    | 3   | 320.654mb +0.00% | 1.288s -13.57%   | ±2.10% +9.01%   |
| TextLoaderBench    | bench_load_10k | 1    | 3   | 17.739mb +0.00%  | 41.777ms +0.98%  | ±0.35% +1.06%   |
+--------------------+----------------+------+-----+------------------+------------------+-----------------+

Building Blocks

+-------------------------+----------------------------+------+-----+------------------+------------------+------------------+
| benchmark               | subject                    | revs | its | mem_peak         | mode             | rstdev           |
+-------------------------+----------------------------+------+-----+------------------+------------------+------------------+
| RowsBench               | bench_chunk_10_on_10k      | 2    | 3   | 76.436mb +0.00%  | 3.251ms +18.04%  | ±2.84% +103.60%  |
| RowsBench               | bench_diff_left_1k_on_10k  | 2    | 3   | 96.229mb +0.00%  | 184.619ms -3.44% | ±0.67% -43.67%   |
| RowsBench               | bench_diff_right_1k_on_10k | 2    | 3   | 74.755mb +0.00%  | 19.148ms -0.95%  | ±2.59% +44.73%   |
| RowsBench               | bench_drop_1k_on_10k       | 2    | 3   | 77.676mb +0.00%  | 1.900ms +12.86%  | ±1.84% +1907.38% |
| RowsBench               | bench_drop_right_1k_on_10k | 2    | 3   | 77.676mb +0.00%  | 1.934ms +14.30%  | ±1.34% +2576.14% |
| RowsBench               | bench_entries_on_10k       | 2    | 3   | 74.789mb +0.00%  | 2.934ms +11.69%  | ±0.04% -98.04%   |
| RowsBench               | bench_filter_on_10k        | 2    | 3   | 75.318mb +0.00%  | 15.582ms +2.71%  | ±1.37% -24.42%   |
| RowsBench               | bench_find_on_10k          | 2    | 3   | 75.318mb +0.00%  | 15.133ms +2.78%  | ±0.54% -64.08%   |
| RowsBench               | bench_find_one_on_10k      | 10   | 3   | 73.220mb +0.00%  | 2.000μs +5.60%   | ±0.00% -100.00%  |
| RowsBench               | bench_first_on_10k         | 10   | 3   | 73.220mb +0.00%  | 0.400μs 0.00%    | ±0.00% 0.00%     |
| RowsBench               | bench_flat_map_on_1k       | 2    | 3   | 86.777mb +0.00%  | 14.136ms +11.21% | ±2.03% -16.84%   |
| RowsBench               | bench_map_on_10k           | 2    | 3   | 116.137mb +0.00% | 65.199ms +4.02%  | ±0.82% -22.35%   |
| RowsBench               | bench_merge_1k_on_10k      | 2    | 3   | 75.838mb +0.00%  | 1.934ms +4.22%   | ±2.31% +213.55%  |
| RowsBench               | bench_partition_by_on_10k  | 2    | 3   | 78.111mb +0.00%  | 35.180ms +3.97%  | ±0.71% -25.19%   |
| RowsBench               | bench_remove_on_10k        | 2    | 3   | 77.939mb +0.00%  | 4.356ms +11.04%  | ±1.38% -18.70%   |
| RowsBench               | bench_sort_asc_on_1k       | 2    | 3   | 73.366mb +0.00%  | 41.362ms +1.89%  | ±0.53% -42.66%   |
| RowsBench               | bench_sort_by_on_1k        | 2    | 3   | 73.366mb +0.00%  | 41.005ms +2.74%  | ±3.11% +36.47%   |
| RowsBench               | bench_sort_desc_on_1k      | 2    | 3   | 73.366mb +0.00%  | 42.849ms +3.79%  | ±2.10% +31.43%   |
| RowsBench               | bench_sort_entries_on_1k   | 2    | 3   | 75.663mb +0.00%  | 7.978ms +5.82%   | ±3.19% +76.81%   |
| RowsBench               | bench_sort_on_1k           | 2    | 3   | 73.220mb +0.00%  | 29.347ms +1.58%  | ±1.51% +107.81%  |
| RowsBench               | bench_take_1k_on_10k       | 10   | 3   | 73.220mb +0.00%  | 13.830μs +4.13%  | ±1.69% +57.52%   |
| RowsBench               | bench_take_right_1k_on_10k | 10   | 3   | 73.220mb +0.00%  | 16.606μs +4.44%  | ±0.28% -72.41%   |
| RowsBench               | bench_unique_on_1k         | 2    | 3   | 96.231mb +0.00%  | 181.672ms -3.34% | ±0.47% -67.11%   |
| TypeDetectorBench       | bench_type_detector        | 1    | 3   | 59.381mb +0.00%  | 335.529ms -0.70% | ±1.77% +70.95%   |
| TypeDetectorBench       | bench_type_detector        | 1    | 3   | 14.304mb +0.00%  | 68.289ms +2.71%  | ±1.20% -20.52%   |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 115.979mb +0.00% | 381.116ms -3.71% | ±0.64% -61.86%   |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 59.697mb +0.00%  | 191.766ms -0.59% | ±0.89% -71.14%   |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 14.822mb +0.00%  | 41.873ms +2.07%  | ±1.19% -24.52%   |
+-------------------------+----------------------------+------+-----+------------------+------------------+------------------+