I am trying to create a plot from an xarray dataset (ds) with the data grouped by dates using hvplot. ds contains a coordinate ‘time’ of datetime64 and when trying to plot the data like this:
plot = ds.hvplot(x='time', groupby='time.date')
the plot is always empty, it seems like it looks for data that matches the exact datetime (at 0:00:00 of the respective date), although I want to plot all data corresponding to this date.
When grouping like this:
plot = ds.hvplot(x='time', groupby=['time.year', 'time.month', 'time.day'])
it works, however if I do not have contiguous data, it is a pain to adjust the sliders to get a combination where there actually exists data.
I think there has to be a better solution, however I could not find anything.
I would appreciate any help. Thanks in advance!
I do this:
df.assign(date=df.index.strftime('%Y-%m-%d')) \
.hvplot.line(groupby='date',
widget_type='scrubber',
widget_location='bottom',
height=500)
The idea is to temporarily add a new string column based on the date of the DatetimeIndex that can be used for ‘groupby’. It looks quit nice and you can scrub through the days effortlessly.
2 Likes
Thanks for the hint, this is basically what I was looking for. It is almost the same when working with xarray. You can assign an extra coordinate date and group by this coordinate. E.g,
ds.assign_coords(date=ds.time.dt.strftime('%Y-%m-%d')) \
.hvplot.line(x='time', groupby='date', height=500)
1 Like