5 ms·
Good question. I did have a situation where looping over a fairly large Pandas dataframe took 12 minutes, but using the underlying numpy values took only 3 seco
by goatlover 9y ago
Good question. I did have a situation where looping over a fairly large Pandas dataframe took 12 minutes, but using the underlying numpy values took only 3 seconds. The looping was required to do complicated row and column operations.
Similarly, naybe the C++ code was using an inefficient data structure for a large dataset compared to using something like Numpy.