I have taken the right log. Somehow the import command (which would have been the last action in the log) did not start until Slicer crashed and is apparently also not represented in the log. In the QIN lung database it worked slow, but fine (as the log shows).
I tried importing the files into microDicom, which worked fine and took about 4 minutes for the whole 850 images. Then I exported them into a new DICOM database and tried importing that into Slicer. That worked, but microDicom disfigured the headers, so the data was unusable.
After that I tried splitting the database into small packages of about 50 images each, and that finally did the trick. Slicer still becomes irresponsive for the time of importing and is incredibly slow (cumulated time ~7h for ~800 images, link only), but at least it works.
I would argue this is opposite from “it works” … At the same time, if a user considers Slicer is that valuable that it is worth the wait of 7 hours to import images - it speaks volumes about Slicer importance!
I would suggest a screenshare to debug this, but I don’t think I know enough about Windows to be of much help.
Can you send the log of one of these sessions? My experience is that Slicer can import a few hundred slices per minute. You seem to have magnitudes slower speed than that which should be possible to fix. My suspicion is that some fields might be malformed and a lot of warning/error messages are printed, which slows down the process. Getting the log file could confirm this or indicate what other problems might be with this data set.
In a way, apart from the slow import it does everything is supposed to, and I don’t need to import images that often. So, yes, worth the tradeoff.
I don’t know if a screenshare will be possible. We work with identified patient data, and I would probably need to delete the database before I can make sure there’s no privacy concerns. You can
probably see why I’m reluctant to do that. If somebody is willing and able to help, I can discuss it with my supervisor. Or maybe we can find a workaround for that.
As for the log files, The successful import logs are no longer in the “recent logs” list; is there a way to access older logs or are they deleted? If so, I’ll make sure to grab a log the next time
I import data.
We recognised that the problem persisted when we tried to import the dataset into slicer on another windows pc, but worked fine on our old Imac. So I too assume that the dataset itself is a source
of error. We exported the DICOM database through Osirix on our Imac using default options. I know that expertise on two platforms and programs is a lot to ask for, but maybe there is an option we need to check/uncheck?
Only the last 10 or so log files are preserved, so you need to re-import to generate the logs again (unless you can find the deleted files in your recycling bin).
.