search

man-group / dtale

Visualizer for pandas data structures
http://alphatechadmin.pythonanywhere.com
GNU Lesser General Public License v2.1
4.78k stars 405 forks source link

Heatmaps break locked columns #800

Closed ms32035 closed 1 year ago

ms32035 commented 1 year ago

In my dataframe of several columns I have 5 locked columns (3 text, and 2 numeric). When I switch to highlight to heatmap (by columns or overall) the locked text columns are hidden, and as a result I see 5 numeric columns as locked columns. I guess the correct way would be to not hide locked columns in the heatmap view

aschonfeld commented 1 year ago

So the reason for certain columns being hidden is that they can't correctly be used in the heatmap calculation. IF you want to display everything (anything that can't be calculation correctly just gets displayed in gray background) you can click this option: Screenshot 2023-09-13 at 12 23 00 PM

I can certainly look into the scenario where you haven't clicked that option and you have columns locked that can't be included in the heatmap calculation. It appears that those column disappear and then it hides the next unlocked columns it's place.

ms32035 commented 1 year ago

Thanks, the latter is exactly the scenario I am describing

aschonfeld commented 1 year ago

@ms32035 just release v3.5.0 to pypi (will be on conda-forge shortly), which includes this fix. Let me know if you have any issues

ms32035 commented 1 year ago

The behaviour changed in that now only the numeric columns remain locked, so the number of locked columns goes down from 5 to 2. Not sure if that's a desired/correct behavior, but I guess that's up to you to decided. Either way, thanks for looking at it so quickly :rocket:

aschonfeld commented 1 year ago

I don't really know of any other option for behavior. If you select the "Show All Heatmap Columns" option then all 5 locked columns will remain. So in this instance it was the "desired" behavior.