Hi Franz - Which of the files that you sent us are the original files (the bvalfile and bvecfile from your dmrirc)? I could not find any files that had 33 lines among your attachments so I'm not sure what your input files look like.
Thanks, a.y
On Tue, 5 Jul 2011, Franz Liem wrote:
Hello again.
I dug into the code and could locate the problem (I don't have a solution, though. My coding skills are quite amateurish.) I think the problem somehow lies in the format of my bvecs file.
flip4fsl is not able to read it correctly:
if (-e $inbvecs) then echo "INFO: found $inbvecs, converting to FSL format" if $fslflipx then set sign = `echo $inorient | sed "s/[RAS]/+\ /g; s/[LPI]/-\ /g"` endif cp /dev/null $outbvecs foreach j (1 2 3) foreach k (1 2 3) if ($order[$k] == $j) then printf '%s ' `cat $inbvecs | awk -v sgn=$sign[$k]1 -v n=$k '{print sgn*$n}'` \ >> $outbvecs printf '\n' >> $outbvecs endif end end endif
I tried awk '{ print $1 }' bvecs (with my original bvecs file) this printed: 0
This has something to do with the file's carriage return. I then produced a new bvecs file by manually typing values into a new text file. Now awk '{ print $1 }' bvecs delivers one column with 33 lines.
Nevertheless, the bvecs output of flip4fsl is weird (full of zeros; 3 x 66; see attachment). I don't really get how this printf & awk command works.
Any comments are very much appreciated. Franz