I am new to vaex. Just started using it to speed up some groupby + agg.nunique operations on ~40 million rows Data Frame in jupyter notebook. It works much faster than pandas, I am really excited to use it more often but sometimes I experience weird error: executing some simple vaex notebook cell with simple filter code like:
vf[vf.Item_count >1]
finishes under 1s but when I run the same cell again it can take several minutes and won't respond to keyboard interruption.
I run vaex and pandas inside VSCode on Win10 machine with 32GB of RAM. My Data Frame in pandas takes around 1GB.
Could you please help me navigate around those slow downs ?
