Indexing Problem, when importing CSV file since Version update

[darbab2s]

I installed vaex a few months ago (like 3 or so). I updated it yesterday.

Now when i import a CSV file via df = vaex.from_csv('assets/annotation_example3.txt',delimiter=";") it results in this:

instead of this:

1dPlots like this work fine: df.plot1d(df.value) and produce:

but when i try to plot a heatmap via display(df[df['sign'] == 'n'].widget.heatmap('value','backvalue', colorbar= True, c = 'Green')) it results in this:

instead of the nice heatmap that i got before updating

what am i doing wrong?

I really like your work, but currently i cant find ways to get my code(more than what is shown here) to work since the update.

[maartenbreddels]

I updated it yesterday.

which versions do you have installed now? (pip list | grep vaex)

[darbab2s]

[darbab2s]

Do you need any further information?

Also i was wondering why i even got vaex 3 via the pip install command in the 4.0 docs and not version 4, which version do i get if i use the github sources?

[maartenbreddels]

because vaex v4 is still an alpha release, you can get it by ...

# being specific
$ pip install vaex==4.0.0a8
# or ask for pre releases
$ pip install vaex --pre

Could you try upgrading first. I still don't get why you see that really old heatmap though with vaex-jupyter 0.5.2

[darbab2s]

upgrading just resulted in a lot of Requirements already satisfied, nothing new.

[maartenbreddels]

but, do you still have vaex 3?

[darbab2s]

same as before

[maartenbreddels]

pip install vaex==4.0.0a8 should change that.

[darbab2s]

Now i´m getting an ImportError: DLL load failed: on import vaex, the advised solution seems to be reinstalling Python, which version would you recommend?

[JovanVeljanoski]

Have you tried re-installing vaex recently? There have been several stable version released since this issue was created.