Requests: content-length is missing from requests

[cuzzasoft]

Can reproduce with the below:

import requests
r = requests.get('https://downloads.atlassian.com/software/confluence/downloads/atlassian-confluence-6.0.5.zip', stream=True)
print r.headers.get('content-length')
None

This results in None - I'm using content-length for a download progress bar and can't get it with requests ;(

If we use something else like urllib it returns it:

import urllib
r = urllib.urlopen(url).info()
print r
Content-Type: application/zip
Content-Length: 541131731

This used to work a while back, not sure if anything changed. Reproduced with Python 2.7.13 & 3.6.0 with requests 2.13.0.

[sigmavirus24]

If you look at the headers that are returned, you will notice

Transfer-Encoding: chunked

If you compare the headers that are sent, however, using httpbin, you might notice that Requests also sends Accept-Encoding: gzip, deflate. If you remove that header, e.g.,

r = requests.get(url, stream=True, headers={'Accept-Encoding': None})

Then you get your Content-Length header. This has been a header that's been set for quite a long time (I think since around when 1.0 was released) so you must have last used this years ago. I hope this helps.

Cheers!
Ian

[cuzzasoft]

@sigmavirus24 you're the best, thanks - I totally missed this. 👍

[quyleanh]

@sigmavirus24 sorry for mention. I still have this issue even when using your suggest.

for example when I want to get content-length from this image link

https://kenhsinhvien.vn/topic/conan-chap-699-tieng-viet.371791/ there is no result

Could you please give some advice?

[sigmavirus24]

@quyleanh it's possible that server doesn't send the content length ever. Some servers prefer not to calculate that if possible. Regardless, this isn't a question and answer support forum and you'll have better luck on stack overflow

[quyleanh]

@sigmavirus24 thank you for replying. Actually I have already asked on stackoverflow.
The thing is although I cannot get the Content-Length with get method, I still can get with head method. So the server sent the content length, right? Then what is the problem?
You can try with the link I send above.

[eziolotta]

You need to add Content-Encoding param , in my case only with this works, and content-length is not zero

r = requests.get(url, stream=True, headers={'Accept-Encoding': None, 'Content-Encoding':'gzip' })