Adding rows manually to StreamingHttpResponse (Django)

Question

I am using Django’s StreamingHttpResponse to stream a large CSV file on the fly. According to the docs, an iterator is passed to the response’s streaming_content parameter:

import csv
from django.http import StreamingHttpResponse

def get_headers():
    return ['field1', 'field2', 'field3']

def get_data(item):
    return {
        'field1': item.field1,
        'field2': item.field2,
        'field3': item.field3,
    }

# StreamingHttpResponse requires a File-like class that has a 'write' method
class Echo(object):
    def write(self, value):
        return value


def get_response(queryset):
    writer = csv.DictWriter(Echo(), fieldnames=get_headers())
    writer.writeheader() # this line does not work

    response = StreamingHttpResponse(
        # the iterator
        streaming_content=(writer.writerow(get_data(item)) for item in queryset),
        content_type='text/csv',
    )
    response['Content-Disposition'] = 'attachment;filename=items.csv'

    return response

My question is: how can I manually write a row on the CSV writer? manually calling writer.writerow(data) or writer.writeheader() (which also internally calls writerow()) does not seem to write to the dataset, and instead only the generated / streamed data from streaming_content is written on the output dataset.

Asked By: silentbugs

||

Source

Answer 1

The answer is yielding results with a generator function instead of calculating them on the fly (within StreamingHttpResponse’s streaming_content argument) and using the pseudo buffer we created (Echo Class) in order to write a row to the response:

import csv
from django.http import StreamingHttpResponse

def get_headers():
    return ['field1', 'field2', 'field3']

def get_data(item):
    return {
        'field1': item.field1,
        'field2': item.field2,
        'field3': item.field3,
    }

# StreamingHttpResponse requires a File-like class that has a 'write' method
class Echo(object):
    def write(self, value):
        return value

def iter_items(items, pseudo_buffer):
    writer = csv.DictWriter(pseudo_buffer, fieldnames=get_headers())
    yield pseudo_buffer.write(get_headers())

    for item in items:
        yield writer.writerow(get_data(item))

def get_response(queryset):
    response = StreamingHttpResponse(
        streaming_content=(iter_items(queryset, Echo())),
        content_type='text/csv',
    )
    response['Content-Disposition'] = 'attachment;filename=items.csv'
    return response

Answered By: silentbugs

Answer 2

The proposed solution can actually lead to incorrect/mismatched CSVs (header mismatched with data). You’d want to replace the affected section with something like:

header = dict(zip(fieldnames, fieldnames))
yield writer.writerow(header)

instead. This is from the implementation of writeheader https://github.com/python/cpython/blob/08045391a7aa87d4fbd3e8ef4c852c2fa4e81a8a/Lib/csv.py#L141:L143

For some reason, it’s not behaving well with yield

Hope this helps someone in the future 🙂

Also note that no fix is needed if using python 3.8+ because of this PR: https://bugs.python.org/issue27497

Answered By: tr33hous

Answer 3

you can chain generator using itertools in python to add header row to the queryset row

here is how you do it:

import itertools

def some_streaming_csv_view(request):
    """A view that streams a large CSV file."""
    # Generate a sequence of rows. The range is based on the maximum number of
    # rows that can be handled by a single sheet in most spreadsheet
    # applications.
    headers = [["title 1", "title 2"], ]
    row_titles = (header for header in headers) # title generator

    items = Item.objects.all()
    rows = (["Row {}".format(item.pk), str(item.pk)] for item in items)
    pseudo_buffer = Echo()
    writer = csv.writer(pseudo_buffer)
    rows = itertools.chain(row_titles, rows)  # merge 2 generators
    return StreamingHttpResponse(
        (writer.writerow(row) for row in rows),
        content_type="text/csv",
        headers={'Content-Disposition': 'attachment; filename="somefilename.csv"'},
    )

and you will get csv with the title and the queryset:

title 1, title 2
1, 1
2, 2
...

Answered By: Linh Nguyen

Adding rows manually to StreamingHttpResponse (Django)

Question:

Answers: