group_label keyword doesn't change group name

[ahuang11]

tst = (hv.Curve([0, 1, 2], group='agroup', label='alabel') + hv.Curve([3, 4, 5], group='bgroup', label='blabel'))
print(tst)

print('\nhvplot')

tst_df = pd.DataFrame({'x1': [0, 1, 2], 'x2': [3, 4, 5]})
print(tst_df.hvplot('index', 'x1', label='alabel', group_label='agroup') + tst_df.hvplot('index', 'x2', label='blabel', group_label='bgroup'))

[philippjfr]

The fact that both have the name group in them is somewhat confusing but the group_label really has nothing to do with the Element.group. The group_label is used to give the dimension that is created when plotting a wide dataframe. For example:

Or if you want to think of it in terms of the arguments to pd.melt, it's basically equivalent to the var_name argument.

[ahuang11]

Is there another way to label groups in hvplot besides using the .regroup method? I think when I first tried passing in group='something', it said the nearest kwarg was group_label, so I had assumed that group_label was the equivalent of group in vanilla HoloViews.

[philippjfr]

I wouldn't be opposed to pass group into hvplot directly but currently there is no such mechanism.