Edit page in Livemark
(2022-09-19 18:33)

Stream Scheme

Frictionless supports using data stored as File-Like objects in Python.

Reading Data

It's recommended to open files in byte-mode. If the file is opened in text-mode, Frictionless will try to re-open it in byte-mode.

You can read Stream using Package/Resource, for example:

from pprint import pprint
from frictionless import Resource

with open('table.csv', 'rb') as file:
  resource = Resource(file, format='csv')
  pprint(resource.read_rows())
[{'id': 1, 'name': 'english'}, {'id': 2, 'name': '中国人'}]

Writing Data

A similiar approach can be used for writing:

from frictionless import Resource

source = Resource(data=[['id', 'name'], [1, 'english'], [2, 'german']])
target = source.write(scheme='stream', format='csv')
print(target)
print(target.to_view())
{'name': 'memory',
 'type': 'table',
 'data': [],
 'scheme': 'stream',
 'format': 'csv',
 'mediatype': 'text/csv'}
+----+-----------+
| id | name      |
+====+===========+
|  1 | 'english' |
+----+-----------+
|  2 | 'german'  |
+----+-----------+
It's a beta version of Frictionless Framework (v5). Read Frictionless Framework (v4) docs for a version that is currently installed by default by pip.