liac-arff

2.5.0 verified Fri May 01 auth: no python

liac-arff is a Python library for reading and writing ARFF (Attribute-Relation File Format) files, commonly used in WEKA machine learning toolkits. Current version 2.5.0 is the last to support Python 2. It supports sparse ARFF, data generators, and standard Java escape sequences. Release cadence is irregular, with about 1-2 releases per year.

pip install liac-arff

Common errors

error AttributeError: module 'arff' has no attribute 'load' ↓

cause Attempting to import 'load' directly from arff (e.g., `from arff import load`) but the correct usage is to import the module and call arff.load.

fix

Use import arff then data = arff.load(...)

error TypeError: expected str, bytes or os.PathLike object, not _io.BufferedReader ↓

cause Passing an open file handle where a path string is expected (often from mixing with other libraries).

fix

Ensure you pass the file object to arff.load(), not the path. If a path is needed, open it first: with open(path) as f: arff.load(f)

Warnings

gotcha The `load()` function expects a file-like object, not a file path. Use `open()` or `io.StringIO`. ↓

fix Use `with open('file.arff') as f: data = arff.load(f)`

gotcha When using `return_type='generator'`, the returned object is a generator that yields rows incrementally, but the `attributes` key is only available after the first iteration or after exhausting the generator. Accessing `data['attributes']` before iterating may raise an error or return an empty list. ↓

fix Iterate over the generator first or use `list(data['data'])` to consume it, then access `data['attributes']`.

breaking Python 2.7 support removed in version 2.5.0. Users on Python 2 must stay on version 2.4.0 or earlier. ↓

fix Upgrade to Python 3.5+ or pin liac-arff to <=2.4.0.

Imports

arff
```
import arff
```
load
wrong
```
from arff import load
```
correct
```
data = arff.load(open('file.arff'))
```
load is a function in the arff module, not a top-level import. Import the module first.

Quickstart

Load an ARFF file and access its attributes and data.

import arff

# Load ARFF file
with open('example.arff') as f:
    data = arff.load(f)
    print(data['attributes'])
    print(data['data'])

# Load with generator (lazy loading)
with open('example.arff') as f:
    for row in arff.load(f, return_type='generator'):
        print(row)