Back to the main page.

Bug 1884 - Implement fallback .m files for nan*

Reported 2012-12-12 10:13:00 +0100
Modified 2013-09-18 13:19:12 +0200
Product: FieldTrip
Component: core
Version: unspecified
Hardware: PC
Operating System: Windows
Importance: P3 normal
Assigned to: Robert Oostenveld
Depends on:
See also:

Jörn M. Horschig - 2012-12-12 10:13:00 +0100

see bug 1768

Jörn M. Horschig - 2013-01-23 13:53:50 +0100

nanstd.m nansum.m nanmean.m nanvar.m

Jörn M. Horschig - 2013-01-23 14:01:35 +0100

- make own compiled and native folder - test script for comparing these and testing 'weird' cases like using strings/function handles nanmean = nansum / nannumel; nanstd = sqrt(nanvar);

Jörn M. Horschig - 2013-01-23 14:56:33 +0100

the nannumel mex file does not behave like matlabs numel - instead of returning a single number it returns a numel per matrix index. it would make sense to abolish nannumel and let the nanmex files determining this implicitly, since they have to do this anyway. I committed nansum.m and nansum.m as fallbacks. I tried nanvar_base but did not manage to easily get it done in a one-liner, so I stopped Sending nanmean.m Sending nansum.m Transmitting file data .. Committed revision 7391.

Jörn M. Horschig - 2013-01-23 14:59:14 +0100

maybe we can use these as fallbacks:

Eelke Spaak - 2013-08-21 16:20:34 +0200

This is fixed, as all relevant nan*.m files are present in /external/stats.

Robert Oostenveld - 2013-09-18 13:19:12 +0200

closed multiple bugs that have been resolved