Back to the main page.

Bug 2864 - ft_sourcedescriptives produceses NaNs only (since revision 10263)

Reported 2015-03-08 15:33:00 +0100
Modified 2019-08-10 12:31:07 +0200
Product: FieldTrip
Component: core
Version: unspecified
Hardware: PC
Operating System: Linux
Importance: P5 normal
Assigned to:
Depends on:
See also:

Jens Klinzing - 2015-03-08 15:33:10 +0100

Since rev. 10263 (Friday, Feb 27) I get exclusively NaNs in data.avg.pow when calling ft_sourcedescriptives. Testdata (2 trials, 250MB): data_des = ft_sourcedescriptive([], data) data_des = freq: 1.3977e+01 cumtapcnt: [58x1 double] dim: [20 25 22] filter: {11000x1 cell} inside: [11000x1 logical] pos: [11000x3 double] method: 'rawtrial' df: 58 trialinfo: [58x1 double] cfg: [1x1 struct] avg: [1x1 struct] var: [1x1 struct] sem: [1x1 struct] In the ouput, data_des.avg.pow contains 11000x1 doubles. However, after revision 10263 they are all NaNs.

Robert Oostenveld - 2015-03-08 19:47:31 +0100

I did make a change somewhere around that time related to how the inside/outside indices are handled. Yes, in rev 10263 the allocation was changed into nan(size(pow)), but later the code incrementally adds each trial to the nan. And nan+something is nan, hence it fails. I just fixed it. mac011> svn commit ft_sourcedescriptives.m Sending ft_sourcedescriptives.m Transmitting file data . Committed revision 10269.

Jens Klinzing - 2015-03-09 13:53:41 +0100

Thanks Robert!

Robert Oostenveld - 2019-08-10 12:31:07 +0200

This closes a whole series of bugs that have been resolved (either FIXED/WONTFIX/INVALID) for quite some time. If you disagree, please file a new issue on