python - Parallel processing with Numpy.loadtxt() -
i have >100mb file needs read numpy.loadtxt()
the reading part main bottleneck in code. 72mb file takes 17.3s
is somehow possible read in parallel way file using loadtxt()
if possible without splitting file.
it looks numpy.loadtxt() problem.
http://wesmckinney.com/blog/?p=543
http://codrspace.com/durden/performance-lessons-for-reading-ascii-files-into-numpy-arrays/
according these sites, you're better off not using numpy's load function @ all.
pandas.read_csv , read_table should helpful pandas module.
Comments
Post a Comment