The release candidate of pandas 1.0 is already online, and it might become the next stable release. What are the plans to move DSS into this new version? Could it be any workaround for test it before it becomes supported by the main python environment?
Bringing in new versions of pandas has historically been quite a lot of work because of various internal behavior changes. This is the reason why we're still at 0.23 and not yet at 0.24 which includes a behavior change on integers that would have far-rippling impacts on customer code.
We are looking into ways to be able to incorporate multiple versions of pandas but cannot announce any ETA at the moment.
You can create a new code environment, without the "core packages" and put pandas 1.0, but you won't be able to "import dataiku" in it.
Thanks Clément for the quick reply. I understand that is going to be difficult, but I'm happy to hear you are working towards a way to keep up to date with pandas.
A little off topic, but also related to Dataiku future plans: will you try to use jupyter labs in the future?
I am facing challenges using Pystan>2.14 & Fbprophet>0.6 due to this older version of Pandas=0.23.4 in DSS-7.0.2.
As I have tested within DSS-7.0.2, Pystan & Fbprophet works with Pandas1.1.4
@Suman, pandas 1 started to be supported with DSS 8.0.0 (https://doc.dataiku.com/dss/latest/release_notes/8.0.html#support-for-pandas-1-0). For version 7 you can't have pandas 1 at the same time you have dataiku libraries installed.