Skip to main content

ARFF Format

Read Weka ARFF (Attribute-Relation File Format) datasets as flat attribute dictionaries. ARFF is read-only in this release.

Overview​

PropertyValue
Format idarff
ClassARFFIterable
Extensions.arff
ReadYes
WriteNo
Extraarff (liac-arff)
Maturitystable

File Extensions​

  • .arff — Weka Attribute-Relation File Format

Implementation Details​

Reading​

  • Loads the full file via liac-arff (arff.loads)
  • Attribute names become keys; missing values (?) are None
  • Relation name is included as _relation
  • Sparse ARFF rows are densified before yield
  • read_bulk() uses in-memory slicing for efficiency

Writing​

Writing is not supported. write() / write_bulk() raise WriteNotSupportedError.

Key Features​

  • Dense and sparse ARFF instances
  • Relation metadata via _relation
  • In-memory bulk reads

Usage​

from iterable import open_iterable

with open_iterable("weather.arff") as source:
    for row in source:
        print(row)

Record shape:

{"outlook": "sunny", "temperature": 85, "play": "no", "_relation": "weather"}

Parameters​

ParameterTypeDefaultRequiredDescription
encodingstr'utf8'NoText encoding

Error Handling​

  • ImportError: Missing liac-arff — install with pip install iterabledata[arff] (or pip install liac-arff)
  • WriteNotSupportedError: ARFF write mode is not currently supported
  • FileNotFoundError: Path is wrong or the file is missing
  • Corrupt / invalid ARFF may raise parse errors from liac-arff
  • UnicodeDecodeError: Wrong encoding — set encoding in iterableargs

See Troubleshooting for more help.

Installation​

pip install 'iterabledata[arff]'

Limitations​

  1. Read-only
  2. Full file loaded into memory
  3. Requires liac-arff