2

I have a timeseries of data and I'd like to use VAEX to manipulate it. I need to groupby an integer "species" column and then also bin by minute. I have tried

    air_df.groupby(by = [vaex.BinnerTime(air_df["DT"],resolution = "m"),air_df["species"]])

but it gives an error "pyarrow.lib.ArrowNotImplementedError: Unsupported datetime64 time unit" It can manage grouping on both columns seperately, but when combined it fails. Is there any way to do this?

Martyn Ward
  • 36
  • 1
  • Wondering if you ever found a solution to this problem? – labourday Feb 23 '22 at 20:46
  • Hi, Never found a solution within vaex, but to be honest I stopped trying a long time ago. You can create your own column using the base number of minutes since a time of your choice and then group on that as an integer value, but it's extra work that I don't want, especially when pandas resamples by time effortlessly. You can convert back and forth from vaex to pandas so that's the easiest option. – Martyn Ward Feb 25 '22 at 09:16

0 Answers0