pd.DataFrame not in completion results

[qubitron]

Environment data

• Language Server version: 2020.6.1
• OS and version: Windows 10
• Python version (& distribution if applicable, e.g. Anaconda): Python 3.8.0 32-bit virtualenv

Expected behaviour

pd.DataFrame is in completion results

Actual behaviour

[gokturksm]

Same behavior with pd.Series, pd.Timestamp etc. They are not exactly functions I guess, maybe that's the cause of the problem?

[gramster]

I think the stubs are alright here, but the only class that's showing as a completion off pd. is the public one defined in __init__.py itself, not the ones that are imported from other modules by __init__.py. Those classes work fine once you have them; i.e. you get completions for pd.DataFrame(). etc. So there's something odd going on.

[erictraut]

Is this perhaps related to this bug? microsoft/pyright#782

[gramster]

Looks like it but I will wait to confirm before marking this as fixed.

[Minituff]

I am getting this error as well. For example, Pylance thinks this is invalid because it cannot see the types of DataFrame:
df = pd.DataFrame.from_dict(my_dict)
I get IntelliSense when I use Jedi alone without Pylance.

Pylance Version: 2020.6.1
Python Language Server Version: 2020.6.91350
Python: 3.7.6
OS: Windows 10

[jakebailey]

This issue has been fixed in version 2020.7.0, which we've just released. You can find the changelog here: https://github.com/microsoft/pylance-release/blob/master/CHANGELOG.md#202070-9-july-2020

[Minituff]

That was a fast update!
I just tried out the new 2020.7.0 release. It looks like the errors have stopped, but I am still not getting IntelliSense like I get with Pyright for pandas dataframes. Should I open a new issue for this?

[jakebailey]

Yes, please do. Pyright doesn't ship any pandas stubs, but we do, so there may be a difference there where it's reading the "real" code.