Improve the CSVExtractor by removing duplicated operations#1665
Merged
norberttech merged 1 commit intoflow-php:1.xfrom May 22, 2025
Merged
Improve the CSVExtractor by removing duplicated operations#1665norberttech merged 1 commit intoflow-php:1.xfrom
CSVExtractor by removing duplicated operations#1665norberttech merged 1 commit intoflow-php:1.xfrom
Conversation
Contributor
Flow PHP - BenchmarksResults 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 | 4.776mb -0.01% | 417.596ms -28.25% | ±0.57% -7.90% |
| ExcelExtractorBench | bench_extract_10k_ods | 1 | 3 | 65.486mb +0.00% | 1.044s -2.68% | ±0.65% +170.21% |
| ExcelExtractorBench | bench_extract_10k_xlsx | 1 | 3 | 67.532mb +0.00% | 1.670s -0.38% | ±0.20% -58.57% |
| JsonExtractorBench | bench_extract_10k | 1 | 3 | 5.018mb +0.00% | 1.284s +0.31% | ±2.71% +475.46% |
| ParquetExtractorBench | bench_extract_10k | 1 | 3 | 86.321mb +0.00% | 921.218ms -0.33% | ±0.33% -14.05% |
| TextExtractorBench | bench_extract_10k | 1 | 3 | 4.499mb +0.01% | 38.380ms -1.73% | ±0.26% -76.39% |
| XmlExtractorBench | bench_extract_10k | 1 | 3 | 4.494mb +0.01% | 604.170ms -0.01% | ±0.07% -77.08% |
+-----------------------+------------------------+------+-----+-----------------+-------------------+-----------------+
Transformers+---------------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
| benchmark | subject | revs | its | mem_peak | mode | rstdev |
+---------------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
| RenameEntryTransformerBench | bench_transform_10k_rows | 1 | 3 | 123.236mb +0.00% | 66.514ms +1.53% | ±0.81% +4.55% |
| RenameEachEntryTransformerBench | bench_transform_10k_rows | 1 | 3 | 18.498mb +0.00% | 72.976ms -0.43% | ±0.17% -52.31% |
+---------------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
Loaders+--------------------+----------------+------+-----+------------------+-----------------+----------------+
| benchmark | subject | revs | its | mem_peak | mode | rstdev |
+--------------------+----------------+------+-----+------------------+-----------------+----------------+
| CSVLoaderBench | bench_load_10k | 1 | 3 | 62.435mb -0.00% | 85.168ms -4.23% | ±0.94% +20.56% |
| JsonLoaderBench | bench_load_10k | 1 | 3 | 79.706mb +0.00% | 96.908ms -0.57% | ±1.09% -43.40% |
| ParquetLoaderBench | bench_load_10k | 1 | 3 | 165.387mb +0.00% | 20.705s -0.73% | ±0.10% -69.05% |
| TextLoaderBench | bench_load_10k | 1 | 3 | 17.805mb +0.00% | 30.994ms -2.80% | ±0.31% -0.25% |
+--------------------+----------------+------+-----+------------------+-----------------+----------------+
Building Blocks+-------------------+----------------------------+------+-----+------------------+------------------+-----------------+
| benchmark | subject | revs | its | mem_peak | mode | rstdev |
+-------------------+----------------------------+------+-----+------------------+------------------+-----------------+
| EntryFactoryBench | bench_entry_factory | 1 | 3 | 101.784mb +0.00% | 648.727ms -1.19% | ±0.67% -43.26% |
| EntryFactoryBench | bench_entry_factory | 1 | 3 | 53.134mb +0.00% | 329.581ms +1.98% | ±0.96% +130.07% |
| EntryFactoryBench | bench_entry_factory | 1 | 3 | 14.384mb +0.00% | 68.681ms -5.04% | ±0.66% -76.11% |
| RowsBench | bench_chunk_10_on_10k | 2 | 3 | 93.389mb +0.00% | 3.516ms -5.99% | ±2.91% -16.13% |
| RowsBench | bench_diff_left_1k_on_10k | 2 | 3 | 110.758mb +0.00% | 235.154ms -0.44% | ±0.34% -67.90% |
| RowsBench | bench_diff_right_1k_on_10k | 2 | 3 | 93.478mb +0.00% | 23.450ms -3.43% | ±0.97% -25.93% |
| RowsBench | bench_drop_1k_on_10k | 2 | 3 | 94.264mb +0.00% | 1.689ms +4.65% | ±3.75% +5.03% |
| RowsBench | bench_drop_right_1k_on_10k | 2 | 3 | 94.264mb +0.00% | 1.583ms -6.10% | ±2.93% +30.87% |
| RowsBench | bench_entries_on_10k | 2 | 3 | 92.424mb +0.00% | 3.431ms -6.01% | ±2.23% -1.16% |
| RowsBench | bench_filter_on_10k | 2 | 3 | 92.953mb +0.00% | 16.304ms -2.26% | ±2.97% +138.88% |
| RowsBench | bench_find_on_10k | 2 | 3 | 92.953mb +0.00% | 15.472ms -3.45% | ±0.51% -23.66% |
| RowsBench | bench_find_one_on_10k | 10 | 3 | 91.642mb +0.00% | 2.000μs +0.30% | ±0.00% -100.00% |
| RowsBench | bench_first_on_10k | 10 | 3 | 91.642mb +0.00% | 0.400μs -20.00% | ±0.00% +0.00% |
| RowsBench | bench_flat_map_on_1k | 2 | 3 | 100.703mb +0.00% | 14.648ms -7.26% | ±0.55% -48.40% |
| RowsBench | bench_map_on_10k | 2 | 3 | 130.130mb +0.00% | 67.425ms -3.83% | ±1.01% -19.47% |
| RowsBench | bench_merge_1k_on_10k | 2 | 3 | 93.473mb +0.00% | 1.526ms +1.88% | ±0.61% -78.41% |
| RowsBench | bench_partition_by_on_10k | 2 | 3 | 96.841mb +0.00% | 62.472ms -1.39% | ±0.14% -82.36% |
| RowsBench | bench_remove_on_10k | 2 | 3 | 94.526mb +0.00% | 4.166ms +8.75% | ±2.86% -18.40% |
| RowsBench | bench_sort_asc_on_1k | 2 | 3 | 92.003mb +0.00% | 39.951ms -2.00% | ±0.82% -73.92% |
| RowsBench | bench_sort_by_on_1k | 2 | 3 | 92.004mb +0.00% | 40.135ms +0.25% | ±1.31% -43.94% |
| RowsBench | bench_sort_desc_on_1k | 2 | 3 | 92.003mb +0.00% | 39.377ms -4.71% | ±1.98% +25.67% |
| RowsBench | bench_sort_entries_on_1k | 2 | 3 | 94.085mb +0.00% | 8.285ms +0.49% | ±0.75% +35.89% |
| RowsBench | bench_sort_on_1k | 2 | 3 | 91.835mb +0.00% | 29.668ms -0.04% | ±2.58% +57.62% |
| RowsBench | bench_take_1k_on_10k | 10 | 3 | 91.642mb +0.00% | 14.382μs -2.03% | ±0.99% -14.08% |
| RowsBench | bench_take_right_1k_on_10k | 10 | 3 | 91.642mb +0.00% | 17.290μs +4.49% | ±2.64% -18.17% |
| RowsBench | bench_unique_on_1k | 2 | 3 | 110.759mb +0.00% | 239.544ms +1.49% | ±1.30% +221.91% |
| TypeDetectorBench | bench_type_detector | 1 | 3 | 42.070mb +0.00% | 430.178ms +2.05% | ±0.44% -33.82% |
| TypeDetectorBench | bench_type_detector | 1 | 3 | 11.448mb +0.00% | 85.267ms -0.21% | ±0.83% +57.51% |
+-------------------+----------------------------+------+-----+------------------+------------------+-----------------+
|
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## 1.x #1665 +/- ##
==========================================
- Coverage 82.08% 82.08% -0.01%
==========================================
Files 703 703
Lines 19064 19059 -5
==========================================
- Hits 15649 15644 -5
Misses 3415 3415
🚀 New features to boost your workflow:
|
Member
|
what are the performance benefits of this? |
stloyd
force-pushed
the
improve-csv-extractor
branch
from
60b5542 to
d9d65d8
Compare
Member
Author
|
I would say it's worth considering ~25-30% of performance boost when reading 10k rows. |
Member
|
Ha! It's indeed faster enough to make a difference! 🎉 So here are the results of following benchmark. Code used to generate benchmark dataset<?php
declare(strict_types=1);
use function Flow\ETL\DSL\from_array;
use function Flow\ETL\DSL\data_frame;
use function Flow\ETL\DSL\overwrite;
use function Flow\ETL\Adapter\CSV\to_csv;
use Faker\Factory;
use Flow\ETL\Rows;
include __DIR__ . '/../../../vendor/autoload.php';
$faker = Factory::create();
$skus = [
['sku' => 'SKU_0001', 'name' => 'Product 1', 'price' => $faker->randomFloat(2, 0, 500)],
['sku' => 'SKU_0002', 'name' => 'Product 2', 'price' => $faker->randomFloat(2, 0, 500)],
['sku' => 'SKU_0003', 'name' => 'Product 3', 'price' => $faker->randomFloat(2, 0, 500)],
['sku' => 'SKU_0004', 'name' => 'Product 4', 'price' => $faker->randomFloat(2, 0, 500)],
['sku' => 'SKU_0005', 'name' => 'Product 5', 'price' => $faker->randomFloat(2, 0, 500)],
];
function generateOrders($faker, array $skus, int $count) : \Generator {
for ($i = 0; $i < $count; $i++) {
yield [
'order_id' => $faker->uuid,
'created_at' => $faker->dateTimeThisYear,
'updated_at' => \random_int(0, 1) === 1 ? $faker->dateTimeThisMonth : null,
'discount' => \random_int(0, 1) === 1 ? $faker->randomFloat(2, 0, 50) : null,
'email' => $faker->email,
'customer' => $faker->firstName . ' ' . $faker->lastName,
'address' => [
'street' => $faker->streetAddress,
'city' => $faker->city,
'zip' => $faker->postcode,
'country' => $faker->country,
],
'notes' => \array_map(
static fn($i) => $faker->sentence,
\range(1, $faker->numberBetween(1, 5))
),
'items' => \array_map(
static fn(int $index) => [
'sku' => $skus[$skuIndex = $faker->numberBetween(1, 4)]['sku'],
'quantity' => $faker->numberBetween(1, 10),
'price' => $skus[$skuIndex]['price']
],
\range(1, $faker->numberBetween(1, 4))
),
];
}
}
$ordersSchema = require __DIR__ . '/schema.php';
data_frame()
->read(from_array(generateOrders($faker, $skus, 1_000_000))->withSchema($ordersSchema))
->saveMode(overwrite())
->write(to_csv(__DIR__ . '/dataset/orders.csv'))
->batchSize(10_000)
->run(function (Rows $rows) {
echo "Generated {$rows->count()} rows\n";
});Schema Code<?php
use function Flow\ETL\DSL\schema;
use function Flow\ETL\DSL\uuid_schema;
use function Flow\ETL\DSL\datetime_schema;
use function Flow\ETL\DSL\float_schema;
use function Flow\ETL\DSL\str_schema;
use function Flow\ETL\DSL\struct_schema;
use function Flow\Types\DSL\type_structure;
use function Flow\Types\DSL\type_string;
use function Flow\Types\DSL\type_list;
use function Flow\ETL\DSL\list_schema;
use function Flow\Types\DSL\type_integer;
use function Flow\Types\DSL\type_float;
return schema(
uuid_schema('order_id'),
datetime_schema('created_at'),
datetime_schema('updated_at', true),
float_schema('discount', true),
str_schema('email'),
str_schema('customer'),
struct_schema(
'address',
type_structure([
'street' => type_string(),
'city' => type_string(),
'zip' => type_string(),
'country' => type_string(),
])
),
list_schema('notes', type_list(type_string())),
list_schema('items', type_list(
type_structure([
'sku' => type_string(),
'quantity' => type_integer(),
'price' => type_float(),
])
))
);Benchmark Code<?php
declare(strict_types=1);
use function Flow\ETL\DSL\data_frame;
use function Flow\ETL\Adapter\CSV\from_csv;
use Flow\ETL\Monitoring\Memory\Consumption;
include __DIR__ . '/../../../vendor/autoload.php';
$schema = require __DIR__ . '/schema.php';
$memory = new Consumption();
$report = data_frame()
->read(from_csv(__DIR__ . '/dataset/orders.csv')->withSchema($schema))
->run(function() use ($memory) {
$memory->current();
},analyze: true);
echo "Total rows: " . \number_format($report->statistics()->totalRows()) . "\n";
echo "Processing time : {$report->statistics()->executionTime->highResolutionTime->toString()}\n";
echo "Memory Max usage : {$memory->max()->inMb()}Mb\n";Benchmarks executed in nix-shell (the one from monorepo)
With following php.ini```ini date.timezone = UTC max_execution_time = 0 error_reporting = 0 display_errors = Off log_errors = Off opcache.enable = 0 opcache.enable_cli = 0 realpath_cache_size = 0 zend.assertions = -1 max_input_time = 3600 max_input_nesting_level = 64 memory_limit = -1 post_max_size = 200M upload_max_filesize = 150M file_uploads = On max_file_uploads = 20 short_open_tag = off ```To make sure results are stable I executed each benchmark 3 times on each branch. Branch
|
norberttech
approved these changes
May 22, 2025
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Change Log
Added
Fixed
Changed
Removed
Deprecated
Security
Description