pyprql

<!-- [](https://codecov.io/gh/prql/PyPrql) -->

pyprql contains:

• pyprql.pandas_accessor — Pandas integration for PRQL
• pyprql.polars_namespace — Polars integration for PRQL
• pyprql.magic — IPython magic for connecting to databases using %%prql
• pyprql.compile — An export of prqlc's compile function

For docs, check out the pyprql docs, and the PRQL Book.

Installation

pip install pyprql

Or, install with optional dependencies:

pip install pyprql[polars]

Usage

Pandas integration

import pandas as pd
import pyprql.pandas_accessor

df = (...)
results_df = df.prql.query("select {age, name, occupation} | filter age > 21")

Polars integration

import polars as pl
import pyprql.polars_namespace

df = (...)
results_df = df.prql.query("select {age, name, occupation} | filter age > 21")

Jupyter Magic

In [1]: %load_ext pyprql.magic
In [2]: %prql postgresql://user:password@localhost:5432/database
In [3]: %%prql
   ...: from p
   ...: group categoryID (
   ...:   aggregate {average unitPrice}
   ...: )
In [4]: %%prql results <<
   ...: from p
   ...: aggregate {min unitsInStock, max unitsInStock}

Compilation

This library exposes prqlc.compile, so we can simply generate SQL:

import pyprql
pyprql.compile("from artists | select track")
print(pyprql.compile("from artists | select track"))

...returns...

SELECT
  track
FROM
  artists

For context, prqlc in Python is the Python binding for the prqlc Rust crate, so only contains functions for compilation; and this library offers broader python integrations and tooling.

Support

This project was created by @charlie-sanders & @rbpatt2019 and is now maintained by the broader PRQL team.