I have the following problem: I have a dataframe with a datetime index and columns that either contain float or boolean values, like this:
TAG_2552 TAG_2526
timestamp
2018-06-17 14:41:48 50.0 True
2018-06-17 14:41:49 51.0 True
2018-06-17 14:41:50 52.0 True
2018-06-17 14:41:51 49.0 True
2018-06-17 14:41:52 55.0 False
2018-06-17 14:41:53 50.0 False
2018-06-17 14:41:54 48.0 False
2018-06-17 14:41:55 54.0 False
2018-06-17 14:41:56 53.0 True
2018-06-17 14:41:57 50.0 True
I would like to plot this, which is not a problem for the float values. For the booleans, I would like to draw rectangles whenever they are True. I tried to use matplotlib.collections to do so .
I tried
df['datetime']=df.index
collection = collections.BrokenBarHCollection.span_where(df['datetime'].dt.date, ymin=0, ymax=1, where=df['TAG_2526']==True, facecolor='green', alpha=.5)
(and other variants) but this did not work.
It is possible to create a collection using
df['timedelta'] = df['datetime']-df['datetime'][0]
collection2 = collections.BrokenBarHCollection.span_where(myProc9['timedelta'].astype('timedelta64[ns]')/np.timedelta64(1,'s'), ymin=0, ymax=1, where=myProc9['TAG_2503']==True, facecolor='green', alpha=.5)
But as I would like to plot against the original datetime because I would like to share the x-Axis, so the only option I can see is to plot everything against myProc9['timedelta'].astype('timedelta64[ns]')/np.timedelta64(1,'s') and relabel the x-Axis.
I checked different pages about changing the datetime-format, but matplotlib.collections does not seem to take anything, the usual error message is
TypeError: float() argument must be a string or a number, not 'datetime.date'
(or whatever else I tried, for various conversions between datetime formats see here: Converting between datetime, Timestamp and datetime64)
I was wondering if there is a better way of doing this (not use matplotlib.collections) or a different way of calling it?
