Feature/window functions

[norberttech]

Change Log

Added

Fixed

Changed

• Expression is now ScalarFunction
• Aggregator is now AggregationFunction

Removed

• struct reference

Deprecated

Security

---

Description

All expressions are now moved under Function namespace, so we have 3 types of functions:

• ScalarFunction - previously expressions
• AggregationFunctions - previously Aggregators
• WindowFunctions - previously also WindowFunctions

Window functions were completely redesigned to work just like they work in SQL

dens_rank()->over(window()->partitionBy(ref('department'))->orderBy(ref('salary')->desc()))

[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   | 34.753mb -0.01%  | 1.144s -1.48%    | ±1.41% +21.39% |
| CSVExtractorBench     | bench_extract_10k | 1    | 3   | 4.610mb -0.09%   | 301.433ms -0.76% | ±0.31% -21.14% |
| JsonExtractorBench    | bench_extract_10k | 1    | 3   | 4.777mb -0.09%   | 1.383s -0.71%    | ±0.33% -48.97% |
| ParquetExtractorBench | bench_extract_10k | 1    | 3   | 239.482mb -0.00% | 1.575s +0.25%    | ±1.10% +51.14% |
| TextExtractorBench    | bench_extract_10k | 1    | 3   | 4.564mb -0.13%   | 24.170ms -0.45%  | ±0.87% +4.42%  |
| XmlExtractorBench     | bench_extract_10k | 1    | 3   | 4.564mb -0.13%   | 408.920ms +0.79% | ±0.62% -8.31%  |
+-----------------------+-------------------+------+-----+------------------+------------------+----------------+

Transformers

+-----------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
| benchmark                   | subject                  | revs | its | mem_peak         | mode            | rstdev         |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
| RenameEntryTransformerBench | bench_transform_10k_rows | 1    | 3   | 110.254mb -0.00% | 64.839ms -1.13% | ±0.49% -79.64% |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+----------------+

Loaders

+--------------------+----------------+------+-----+------------------+------------------+-----------------+
| benchmark          | subject        | revs | its | mem_peak         | mode             | rstdev          |
+--------------------+----------------+------+-----+------------------+------------------+-----------------+
| AvroLoaderBench    | bench_load_10k | 1    | 3   | 94.734mb -0.00%  | 434.447ms -2.75% | ±0.71% -46.36%  |
| CSVLoaderBench     | bench_load_10k | 1    | 3   | 54.650mb +0.11%  | 70.674ms +0.10%  | ±0.96% +325.53% |
| JsonLoaderBench    | bench_load_10k | 1    | 3   | 105.317mb -0.00% | 55.185ms +0.41%  | ±1.70% -11.15%  |
| ParquetLoaderBench | bench_load_10k | 1    | 3   | 320.789mb -0.00% | 1.395s -1.69%    | ±0.26% -48.98%  |
| TextLoaderBench    | bench_load_10k | 1    | 3   | 17.594mb -0.02%  | 40.336ms -0.86%  | ±0.32% -84.69%  |
+--------------------+----------------+------+-----+------------------+------------------+-----------------+

Building Blocks

+-------------------------+----------------------------+------+-----+------------------+------------------+------------------+
| benchmark               | subject                    | revs | its | mem_peak         | mode             | rstdev           |
+-------------------------+----------------------------+------+-----+------------------+------------------+------------------+
| RowsBench               | bench_chunk_10_on_10k      | 2    | 3   | 76.300mb -0.00%  | 2.488ms +14.33%  | ±0.79% -78.77%   |
| RowsBench               | bench_diff_left_1k_on_10k  | 2    | 3   | 96.092mb -0.01%  | 180.233ms +0.02% | ±0.28% -25.89%   |
| RowsBench               | bench_diff_right_1k_on_10k | 2    | 3   | 74.618mb -0.01%  | 17.990ms -0.23%  | ±0.40% -36.81%   |
| RowsBench               | bench_drop_1k_on_10k       | 2    | 3   | 75.438mb -0.00%  | 1.859ms +15.44%  | ±2.28% -28.43%   |
| RowsBench               | bench_drop_right_1k_on_10k | 2    | 3   | 75.438mb -0.00%  | 1.869ms +13.17%  | ±0.44% -80.63%   |
| RowsBench               | bench_entries_on_10k       | 2    | 3   | 74.652mb -0.01%  | 2.723ms +7.37%   | ±0.83% +13.53%   |
| RowsBench               | bench_filter_on_10k        | 2    | 3   | 75.181mb -0.00%  | 15.280ms +6.64%  | ±0.12% -90.99%   |
| RowsBench               | bench_find_on_10k          | 2    | 3   | 75.180mb -0.00%  | 14.838ms +5.89%  | ±1.48% +33.50%   |
| RowsBench               | bench_find_one_on_10k      | 10   | 3   | 73.084mb -0.01%  | 1.794μs -0.67%   | ±2.67% +3.77%    |
| RowsBench               | bench_first_on_10k         | 10   | 3   | 73.084mb -0.01%  | 0.300μs 0.00%    | ±0.00% 0.00%     |
| RowsBench               | bench_flat_map_on_1k       | 2    | 3   | 86.642mb -0.00%  | 13.174ms +1.84%  | ±1.82% +109.39%  |
| RowsBench               | bench_map_on_10k           | 2    | 3   | 115.999mb -0.00% | 64.529ms +0.15%  | ±0.80% +31.28%   |
| RowsBench               | bench_merge_1k_on_10k      | 2    | 3   | 75.701mb -0.00%  | 1.948ms +7.86%   | ±2.42% +108.39%  |
| RowsBench               | bench_partition_by_on_10k  | 2    | 3   | 77.971mb -0.01%  | 34.323ms +7.66%  | ±0.40% +57.44%   |
| RowsBench               | bench_remove_on_10k        | 2    | 3   | 77.802mb -0.00%  | 4.415ms +0.36%   | ±3.19% +587.75%  |
| RowsBench               | bench_sort_asc_on_1k       | 2    | 3   | 73.228mb -0.01%  | 38.189ms +3.53%  | ±0.68% +1026.39% |
| RowsBench               | bench_sort_by_on_1k        | 2    | 3   | 73.229mb -0.01%  | 38.275ms +3.59%  | ±0.21% -26.01%   |
| RowsBench               | bench_sort_desc_on_1k      | 2    | 3   | 73.228mb -0.01%  | 38.654ms +4.43%  | ±0.65% -48.45%   |
| RowsBench               | bench_sort_entries_on_1k   | 2    | 3   | 75.526mb -0.00%  | 7.444ms +1.02%   | ±0.71% +50.93%   |
| RowsBench               | bench_sort_on_1k           | 2    | 3   | 73.083mb -0.01%  | 28.545ms +0.09%  | ±0.49% -38.80%   |
| RowsBench               | bench_take_1k_on_10k       | 10   | 3   | 73.084mb -0.01%  | 13.300μs +2.57%  | ±0.61% -61.06%   |
| RowsBench               | bench_take_right_1k_on_10k | 10   | 3   | 73.084mb -0.01%  | 15.493μs -0.84%  | ±1.97% +64.63%   |
| RowsBench               | bench_unique_on_1k         | 2    | 3   | 96.093mb -0.01%  | 181.981ms -0.13% | ±0.89% +75.98%   |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 115.846mb -0.00% | 757.260ms +0.73% | ±0.50% -65.31%   |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 59.564mb -0.01%  | 386.423ms +0.06% | ±0.42% -28.86%   |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 14.687mb -0.03%  | 79.172ms -0.16%  | ±0.70% +73.78%   |
+-------------------------+----------------------------+------+-----+------------------+------------------+------------------+