Looking up a key in a DataframeGroupBy object is very slow

Question

I group data with DateTime index by Timedelta like so:

compression = datetime.timedelta(minutes=1)
grouped_data = raw_data.groupby([pd.Grouper(freq=compressio