Docs for "pandasutils"?

JustinMinnion

Hello,

My apologies if this is a remedial question, but at the start of every Python recipe the boilerplate code includes an import of:

from dataiku import pandasutils as pdu

Is there documentation for pandasutils? Is it a package that can be used in Python recipes?

I've tried looking through the Dataiku Developer Guide, and I even tried seeing if the dataiku package had its source code published anywhere (I could only find dataiku-api-client-python ), but no luck.

It's actually been (in a funny way) challenging to look for these docs because so many unrelated code samples include this import line.

Any help is appreciated, thanks in advance!
-Justin

Operating system used: OS X

Turribeach

The "dataiku" package is a bit of kludge if you ask me (not to be confused with the dataiku-api-client, see here for Dataiku Python APIs). First of all it isn't available on PyPi but you can install it from a tar bundle you can get from your DSS node. But even that bundle doesn't have all the modules that a full Dataiku installation will have as I noted on this thread. The dataiku module gets imported via sys.path.append(). It lives on the folder where you unpacked the Dataiku installer. In a MacOS installation that will be different since MacOS has a separate installer which leaves the module under:

/Users/[your MacOS ID]/Library/DataScienceStudio/kits/dataiku-dss-[your current DSS version]-osx/python/dataiku

You can find the pandasutils module under ./core/pandasutils.py

JustinMinnion

https://community.dataiku.com/discussion/comment/44750#Comment_44750

Thanks for this reply, @Turribeach ! I was able to grab the tar bundle from my DSS node and see the pandasutils.py source code. It's less exciting than I had hoped but at least I know I'm not missing out on major functionality by not using it.

Thanks again!