there are some hacks: http://mihai.ibanescu.net/chunked-encoding-and-python-requests

hmm any update on this? I am having issues querying large amounts of points. I was considering writing a batch parameter for the client.query() method, but id rather use something that is already implemented.

This is still a problem. :(

Here is a patch that will make chucked work with 1 series of data. I am sure there are plenty of cases where it will not work though.

#45

I'll be taking a look at this soon.

Now that we use python-requests's .json() method. It does not seem to be an issue. If I am wrong, please provide me with an example of a response that does not work.

Thank you for reporting the issue :)

I have realised that my test case only checks for a response with one object in it. I'll re-open this.

Thanks @ReAzem !

@gourneau, @pkaeding, @ilovenwd , @manugarri

I have pushed a fix proposition to branch fix_21_proposition (1a523a4).

What do you guys think?

• It does not add any other library
• It's safer than altering the string (@gourneau's proposition), but surely not faster. Is speed even a concern?

I have not found any better ways of doing this with the standard lib, but this solutions works pretty well, I think.

I opened a PR that is relating to this ticket. When pulling down chunked response its nice for the server side as it utilizes less memory, but client side you'd have to store it all in memory which wouldn't work if you're querying for GBs of data. This PR allows you to stream directly to disk.