Open venaturum opened 2 years ago
The implementation of Stairs.layer requires input vectors to be "grouped by", in case there are repeated values in start and end parameters. This is probably slow for very large datasets. A cython implementation may yield much better results?
Not sure cythonising will improve efficiency here, but checking for duplicates before performing groupby might be worthwhile
Describe the solution you'd like
The implementation of Stairs.layer requires input vectors to be "grouped by", in case there are repeated values in start and end parameters. This is probably slow for very large datasets. A cython implementation may yield much better results?