pymsasdax 2023.1001

Run DAX queries against Analysis Services and get Pandas Dataframes

About

I wanted to consume some info from Azure Analysis Services from python and didn't see a convenient way to to do, so I wrote this. It should also work just fine with XMLA endpoints on Power BI Premium.

pymsasdax is a small Python Module for running DAX queries against Microsoft Analysis Services, using COM Interop. It does some basic typesniffing and returns a best guess Pandas Dataframe.

This does assume that the MSOLAP client is installed - you can get it from here

I've done very little testing, so consider this alpha code. If you run into timeouts, make sure you're setting the timeout to an appropriate duration when creating the Connection.

tidy_column_names will remove brackets and replace spaces with underscores in the returned dataframe's columns. Set it to False in the Connection init if you don't want this behavior.

Also, this is my first module up on pypi and I'm not exactly an expert on python, so feel free to submit an issue or a pull request. If I ended up reinventing the wheel here (ha!) and there was an easier way to do this, also please let me know.

I hope you find this useful!

Python before 3.9

This should actually work fine with for python 3 under 3.9. I've used this code for a couple of years now without incident -- I was just lazy when building this package. I think you'd need backports to support dateparser.

Usage examples

Have an interactive prompt for Login to the resource

from pymsasdax import dax

with dax.Connection(
        data_source='asazure://<region name>.asazure.windows.net/<instance here>,
        initial_catalog='<my tabular database>'
    ) as conn:
    df = conn.query('EVALUATE ROW("a", 1)')
    print(df)

Use an app id

from pymsasdax import dax

with dax.Connection(
        data_source='asazure://<region name>.asazure.windows.net/<instance here>,
        initial_catalog='<my tabular database>'
        uid='app:<client id>@<tenant id>',
        password='<client secret>
    ) as conn:
    df = conn.query('EVALUATE ROW("a", 1)')
    df.to_csv("raw_data.csv", index=False)        

Dev Notes

Current Version - 2023.1001-alpha

Tests

Yes. There aren't any. Feel free to submit a PR.

Building

This might not be right but if you ever go to update pypi -

pip-compile pyproject.toml
bumpver update --minor
python -m pip install -e . 
#test
python -m build
twine check dist/*
twine upload -r pypy dist/*