Back to the main page.
Bug 1622 - ft_fetch_data always gives error when trials overlap in data
| Status | CLOSED FIXED |
| Reported | 2012-08-01 15:07:00 +0200 |
| Modified | 2013-01-17 10:40:35 +0100 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Windows |
| Importance: | P3 normal |
| Assigned to: | Roemer van der Meij |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Roemer van der Meij - 2012-08-01 15:07:53 +0200
Jan-Mathijs Schoffelen - 2012-08-01 15:11:15 +0200
this is appropriate behavior, isn't it? Otherwise it's ambiguous as to from which trial to select the requested samples.
Roemer van der Meij - 2012-08-01 15:13:20 +0200
For now, I added a keyvalue-pair (allowoverlap = true) to fetch_data to escape the error for use in ft_databrowser. The need for this arose in databrowser, where there are overlapping trials and cfg.continous = 'yes'. When zoomed out enough, the requested data is present in several trials. In effect, the samples requested are taken from the last trial where they occur. The only issue I see with allowing overlapping trials in fetch_data is filter-ringing or other preprocessing steps that can affect each sample in a trial differently.
Roemer van der Meij - 2012-08-01 15:18:15 +0200
Ah, wrote the comment in the mean-time. I saw two options: 1) knock out overlapping trials when data was handed as input (it should work fine when read in from disk). This creates different behavior for data on disk vs already read-in. 2) allow it for data that is already read-in. I have a slight preference for the last as it is cleaner coding, but smudgier data-handling. But since it is only for browsing, that was okay in my opinion. But not a very strong preference (the error was just in the way).
Roemer van der Meij - 2012-08-02 15:44:14 +0200
I'll change it to a warning when the exception is used, that should be good enough until someone complains :)
Roemer van der Meij - 2012-08-02 15:48:12 +0200
Changed to 'fixed'