py-r-fst 0.1.0

Python wrapper for the R fst package

py-fst

Python wrapper for the R fst package. The fst package provides a fast, easy and flexible way to serialize data frames.

Installation

From PyPI (once published):

pip install py-fst

From GitHub:

pip install git+https://github.com/msdavid/py-fst.git

From a local directory:

# Standard installation
pip install /path/to/py-fst

# Development mode (changes to code reflect immediately without reinstalling)
pip install -e /path/to/py-fst

Requirements

• Python 3.6+
• R with the 'fst' package installed
• rpy2
• pandas

You need to have R installed and the fst package:

install.packages("fst")

Usage

import pandas as pd
from pyfst import read_fst, write_fst

# Create a sample DataFrame
df = pd.DataFrame({
    'a': [1, 2, 3],
    'b': ['x', 'y', 'z']
})

# Write to fst file
write_fst(df, 'data.fst', compress=50)

# Read from fst file
df_read = read_fst('data.fst')
print(df_read)

# Read only specific columns
df_partial = read_fst('data.fst', columns=['a'])
print(df_partial)

# Read a subset of rows
df_rows = read_fst('data.fst', from_=0, to=1)
print(df_rows)

API

read_fst(path, columns=None, from_=None, to=None, as_data_table=False)

Read a fst file into a pandas DataFrame.

Parameters:

• path: Path to the fst file.
• columns: List of column names to read. If None, all columns are read.
• from_: First row to read (0-based indexing). If None, start from the first row.
• to: Last row to read (0-based indexing). If None, read to the end.
• as_data_table: If True, return a data.table object. Otherwise, return a pandas DataFrame.

Returns:

• pandas.DataFrame or data.table: Data from the fst file.

write_fst(df, path, compress=50)

Write a pandas DataFrame to a fst file.

Parameters:

• df: pandas.DataFrame, data to write.
• path: Path where to save the fst file.
• compress: Compression level (0-100). Higher means more compression but slower.

Returns:

• None

License

Apache 2.0