CSV Tools for Django REST Framework
$ pip install djangorestframework-csv
- For compatibility with Python 3 and Django < 3.2, use djangorestframework-csv<3.0.1
- For compatibility with Python 2.7, use djangorestframework-csv<3
views.py
from rest_framework.views import APIView
from rest_framework.settings import api_settings
from rest_framework_csv import renderers as r
class MyView (APIView):
renderer_classes = (r.CSVRenderer, ) + tuple(api_settings.DEFAULT_RENDERER_CLASSES)
...
Alternatively, to set CSV as a default rendered format, add the following to the settings.py file:
REST_FRAMEWORK = {
# specifying the renderers
'DEFAULT_RENDERER_CLASSES': (
'rest_framework_csv.renderers.CSVRenderer',
),
}
By default, a CSVRenderer
will output fields in sorted order. To specify
an alternative field ordering you can override the header
attribute. There
are two ways to do this:
Create a new renderer class and override the
header
attribute directly:class MyUserRenderer (CSVRenderer): header = ['first', 'last', 'email'] @api_view(['GET']) @renderer_classes((MyUserRenderer,)) def my_view(request): users = User.objects.filter(is_active=True) content = [{'first': user.first_name, 'last': user.last_name, 'email': user.email} for user in users] return Response(content)
Use the
renderer_context
to override the field ordering on the fly:class MyView (APIView): renderer_classes = [CSVRenderer] def get_renderer_context(self): context = super().get_renderer_context() context['header'] = ( self.request.GET['fields'].split(',') if 'fields' in self.request.GET else None) return context ...
Custom labels can be applied to the CSVRenderer
using the labels
dict
attribute where each key corresponds to the header and the value corresponds
to the custom label for that header.
1) Create a new renderer class and override the header
and labels
attribute directly:
class MyBazRenderer (CSVRenderer): header = ['foo.bar'] labels = { 'foo.bar': 'baz' }
Using the renderer with paginated data is also possible with the new PaginatedCSVRenderer class and should be used with views that paginate data
For more information about using renderers with Django REST Framework, see the API Guide or the Tutorial.
To run the tests against the current environment:
$ ./manage.py test
- Fix the project long description metadata
- Remove Django tests for all but the currently supported Django versions (3.2, 4.1, 4.2). Also only keep tests for Python versions compatible with supported Django versions (3.8 - 3.11).
- Removed compatibility with Python 2
- Added compatibility with Django up to 3.2 and 4.2 (thanks to work from @TTycho, @awais786, @arpitjain799, @tirkarthi)
- Add support for byte order markers (BOM) (thanks @Yaoxin)
- Documentation updates (thanks @rjemanuele and @coreyshirk)
- CSVs with no data still output header labels (thanks @travisbloom)
- Include a paginated renderer as part of the app (thanks @masterfloda)
- Generators can be used as data sources for CSVStreamingRenderer (thanks @jrzerr)
- Support for non UTF-8 encoding parsing (thanks @weasellin)
- Make CSVRenderer.render return bytes, and CSVParser.parse expect a byte stream.
- Have data-less renders print header row, if header is explicitly supplied
- Drop Django 1.7 tests and add Django 1.10 tests
- have CSVRenderer.tableize act as a generator when possible (i.e., when a header is explicitly specified).
- Add docs for labels thanks to @radyz
- Fix header rendering in CSVStreamingRenderer thanks to @radialnash
- Improve unicode handling, thanks to @brandonrobertz
- Add support for changing field labels in the
CSVRenderer
, thanks to @soby - Add support for setting
CSVRenderer
headers, labels, and writer_opts asrenderer_context
parameters. - Renamed
CSVRenderer.headers
toCSVRenderer.header
; old spelling is still available for backwards compatibility, but may be removed in the future.
- Support streaming CSV rendering, via @ivancrneto
- Improved test configuration and project metadata, via @ticosax
- Support unicode CSV parsing, and universal newlines, with thanks to @brocksamson
- Renderer handles case where data is not a list by wrapping data in a list, via pull request from @dougvk
- Better cross Python version support, via @paurullan and @vishen
- Support for Python 3, derived from work by @samdobson
- Support consistent ordering of fields in rendered CSV; thanks to @robguttman
- Support specifying particular fields/headers in custom CSV renderer by
overriding the
headers
attribute.
- Support simple CSV parsing; thanks to @sebastibe
- Add the package manifest
- Initial release