phpstats / phibble

v0.7.0 2026-11-01 23:48 UTC

Requires

• php: >=8.3 <8.6
• ext-mbstring: *
• ext-pdo: *
• symfony/polyfill-php84: ^1.31

Requires (Dev)

• illuminate/database: ^12.0
• pestphp/pest: ^4.0
• phpstats/phiz: dev-main
• squizlabs/php_codesniffer: ^3.13

Suggests

• ext-pdo_sqlite: Used by the test suite for the PDO source
• ext-phit: Native CSV reading with Phit\read_csv (phit 0.1.1 or later) and Phit\CsvReader for chunks (phit 0.2.0 or later), from phpstats/phit
• ext-php_polars: The Polars backend (Phibble\Polars\PolarsTable), from phpstats/phpolars
• illuminate/database: Laravel integration: Collection::toDataFrame() and query builder ->phibble() / ->toDataFrame() macros (auto-discovered)

MIT

Published on Packagist.org on 2026-11-01 23:49 UTC

README

phibble is a data frame library for PHP, following R's tibble and readr. A DataFrame holds typed columns, treats null as NA, and returns a new table from every operation. It is the shared table type for phplyr, phit, phpolars, phiz, and CmdStanPHP.

Documentation: https://phibble.phpstats.org/docs

Install

composer require phpstats/phibble

phibble supports PHP 8.3, 8.4, and 8.5, and CI tests all three. It requires the mbstring and PDO extensions. On PHP 8.5, pipelines can also use the |> operator.

Example

In a new directory, install phibble:

composer require phpstats/phibble

Then save this script as summary.php and run php summary.php. It reads the Palmer penguins data from the PHPstats datasets:

<?php

require __DIR__.'/vendor/autoload.php';

use Phibble\DataFrame;

use function Phibble\Expr\{col, mean, n};

$penguins = DataFrame::fromCsv(
    'https://raw.githubusercontent.com/phpstats/datasets/main/csv/penguins.csv'
);

echo $penguins
    ->filter(col('body_mass_g')->isNa()->not())
    ->groupBy(['species'])
    ->summarize(['n' => n(), 'mass' => mean(col('body_mass_g'))]), "\n";

# A tibble: 3 × 3
  species       n  mass
  <chr>     <int> <dbl>
1 Adelie      151 3701.
2 Chinstrap    68 3733.
3 Gentoo      123 5076.

Features

• tibble: tibble(), tribble(), as_tibble(), addRow(), addColumn(), enframe(), deframe(), name repair, strict column access, and printing and glimpse() that match tibble's output.
• Types: float, int, bool, string, factor (with ordered factors and the forcats functions), date, datetime, time, and list-columns. Casts follow R's as.*() rules.
• Input and output: CSV with readr's options and type guessing (and the phit extension's native reader when it is loaded), JSON, PDO queries, Laravel and Doctrine query builders, and Laravel collections.
• Table methods: select, filter, mutate, arrange, groupBy, summarize, distinct, joins (including join_by comparisons and rolling joins), nest, unnest, and tidyr's spec-based pivots, with dplyr 1.1 semantics.
• Expressions: Phibble\Expr describes computations as inspectable, serializable trees that follow dplyr's semantics, with stringr and lubridate basics.
• SQL tables: SqlTable compiles table methods to SQL for SQLite, PostgreSQL, and MySQL, like dbplyr. Anything SQL cannot compute exactly runs in memory.
• Laravel: ->toDataFrame() and ->phibble() on queries and collections.
• Backends: DataFrame (in memory), SqlTable (databases), and PolarsTable (Polars, with the php_polars extension from phpolars) are all part of phibble and implement one Table contract, so the same pipeline runs on each.

Development

composer install
vendor/bin/pest

See the development guide for the R comparison scripts, database tests, and documentation build.