datafusion-cli not installed

[l1t1]

Describe the bug

https://arrow.apache.org/datafusion/user-guide/cli.html
the CLI cannot run

datafusion-cli
-bash: datafusion-cli: command not found

To Reproduce

pip install datafusion

datafusion-cli

Expected behavior

the CLI runs

Additional context

No response

[l1t1]

python module works

python3
Python 3.10.12 (main, Jun 11 2023, 05:26:28) [GCC 11.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import datafusion                                                                          
>>> datafusion.__version__
'35.0.0'

[viirya]

Hmm, I think this issue may move to https://github.com/apache/arrow-datafusion-python?

[l1t1]

thanks.
I aslo post it there
but I also saw the source code in
https://github.com/apache/arrow-datafusion/tree/main/datafusion-cli
is not in
https://github.com/apache/arrow-datafusion-python

[viirya]

datafusion-cli is in this repo. I meant its pypi packaging and release should be done at https://github.com/apache/arrow-datafusion-python.

I think the issue is not datafusion-cli itself or its functionality but the pypi package issue.

[viirya]

Anyway, it is okay to keep it open to get visibility. :)

[alamb]

I think there is a brew package if you are on Mac:

brew install datafusion

[Jefffrey]

Actually, does the datafusion pypi package even include datafusion-cli in the first place? 🤔

I did a quick search through https://github.com/apache/arrow-datafusion-python and found no actual mention of datafusion-cli (though in fairness I'm not familiar with that repo or the process of packaging a Python package)

Maybe the documentation on that CLI user guide page is mistaken on that account? Relevant PR: #8389

cc @Weijun-H

[Weijun-H]

Actually, does the datafusion pypi package even include datafusion-cli in the first place? 🤔

I did a quick search through apache/arrow-datafusion-python and found no actual mention of datafusion-cli (though in fairness I'm not familiar with that repo or the process of packaging a Python package)

Maybe the documentation on that CLI user guide page is mistaken on that account? Relevant PR: #8389

cc @Weijun-H

After checking the documentation in apache/arrow-datafusion-python, I discovered that the current PyPi installation for CLI is incorrect @Jefffrey . Perhaps it's time to implement pip install datafusion-cli 🤔 ?

[SteveLauC]

Perhaps it's time to implement pip install datafusion-cli

I am not sure how DataFusion release procedure works, but if you want to automate it in CI, maturin can help.

I have done this to Topgrade, take a look at this PR if you want to see a real-world example on how it looks like.

[l1t1]

pip install datafusion-cli works now, thanks.