Back to the main page.
Bug 2508 - make it optional whether dimensions are kept after averaging with ft_selectdata
| Status | CLOSED FIXED |
| Reported | 2014-03-21 14:12:00 +0100 |
| Modified | 2014-05-14 20:08:49 +0200 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Mac OS |
| Importance: | P5 normal |
| Assigned to: | Robert Oostenveld |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | 1021 |
| See also: |
Robert Oostenveld - 2014-03-21 14:12:29 +0100
right now ft_selectdata_new is inconsistent with respect to keeping dimensions (freq, chan) or removing them (time, rpt) when averaging. After discussing with Jan-Mathijs (and in the light of the recent discussion at the FT meeting), we consider it worthwile to have both cfg.avgoverxxx = yes/no cfg.keepxxxdim = yes/no to allow dimensions to be retained or removed when averaging.
Robert Oostenveld - 2014-03-21 14:17:33 +0100
note to all: I am now (i.e. this Friday afternoon) working on ft_selectdata_new, so please do not start working on this bug immediately.
Robert Oostenveld - 2014-03-21 14:45:14 +0100
The following commit includes some preliminary changes in the direction indicated for this bug. Around line 133 there is now % keeprptdim = istrue(ft_getopt(cfg, 'keeprptdim', true)); % keepchandim = istrue(ft_getopt(cfg, 'keepchandim', true)); % keepfreqdim = istrue(ft_getopt(cfg, 'keepfreqdim', true)); % keeptimedim = istrue(ft_getopt(cfg, 'keeptimedim', true)); which should be further implemented. svn commit Sending test/test_bug1984_2187.m Sending test/test_ft_selectdata.m Sending utilities/ft_selectdata_new.m Transmitting file data ... Committed revision 9306. r9306 | roboos | 2014-03-21 14:42:23 +0100 (Fri, 21 Mar 2014) | 2 lines bugfix - when working on bug 2509 I detected a problem with the output dimord, which was inconsistent with the remaining fields after averaging. I added a section to the test script for ft_selectdata_new and updated the selectdata function. Furthermore, I implemented a workaround for http://bugzilla.fcdonders.nl/show_bug.cgi?id=2509
Robert Oostenveld - 2014-04-03 11:45:58 +0200
it is now possible to selectively keep/remove dimensions over which data is being averaged. Note that this functionality is not meant to be used in all other FT functions, as it may result in data structures that are not FT compatible (e.g. timelock without time or chan). roboos@mentat001> svn commit test/test_bug2508.m utilities/ft_selectdata_new.m Adding test/test_bug2508.m Sending utilities/ft_selectdata_new.m Transmitting file data .. Committed revision 9340.