将 Pandas DF 绘制为堆积条形图，同时忽略缺失值

Question

I've got a dataframe with workout logs like this:

>>> data
       _date  _week  Distance (km)
0 2020-10-31     44      42.220013
1 2020-10-29     44      10.054135
2 2020-10-25     43      30.103745
3 2020-10-24     43      14.135142
4 2020-10-20     43      10.471132
5 2020-10-17     42      27.164278
6 2020-10-15     42      18.511157

And would like to plot it as a stacked-bar chart per week, like this:

Problem 1: I can't figure out how to pivot or group by the DF before plotting, given that it's got uneven number of workouts per week, up to 4 or 5, but sometimes none at all when I was lazy.

Problem 2: Optionally I'd like to sort them from longest to shortest for aesthetics, so that the longest ones are always the same colour and are at the bottom, etc. But it's nit picking :)

Thanks for any pointers!

Answer 1

You need to create a new field for Run using cumcount and pivot it to columns with week on the index:

df['Run'] = 'Run ' + (df.groupby('_week').cumcount() + 1).astype(str)
df = df.drop('_date', axis=1).set_index('_week').pivot(columns='Run')
df.plot(kind='bar', stacked=True)