[Mne_analysis] Split file saving doesn't work

[Eric Larson]

        External Email - Use Caution        

I suspect I understand why this can happen from the code -- we calculate
the splits using just the size of the actual data, but end up writing that
plus other things (drop log, metadata, etc.) which can push it beyond the
2GB limit. For now you can do something like `split_size='1.9GB'` to work
around the problem, but hopefully we can cook up a proper fix for it soon.

Eric


On Wed, May 6, 2020 at 8:51 AM Alexandre Gramfort <
alexandre.gramfort at inria.fr> wrote:

>         External Email - Use Caution
>
> hi Lau,
>
> please open an issue and make the file available.
>
> you can save it in .h5 format so we can debug the fif IO
>
> Alex
>
> On Wed, May 6, 2020 at 2:13 PM Lau Møller Andersen <lau.andersen at cnru.dk>
> wrote:
> >
> >         External Email - Use Caution
> >
> > Dear list
> >
> > Every now and then I cannot save my epochs objects due to this error.
> This happens only for one of my subjects (out of thirty) and only on one
> particular (bandpass filtered and hilbert transformed) raw file. I have
> tried both the development version and newest stable version.
> >
> > I guess it should take care of it automatically as I didn't change the
> default split size of 2GB of epochs_hilbert.save()
> >
> > All the best
> >
> > Lau
> >
> > _______________________________________________
> > Mne_analysis mailing list
> > Mne_analysis at nmr.mgh.harvard.edu
> > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/mne_analysis
>
> _______________________________________________
> Mne_analysis mailing list
> Mne_analysis at nmr.mgh.harvard.edu
> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/mne_analysis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20200506/7bebae80/attachment.html