Just some useful things that I can never remember:
  • sed -i 's/str1/str2/g' file.ext  (use "" if argument contains a variable)
  • ncrcat files.nc catfile.nc
  • Find string (case-insensitive) in files of specified type (here .f):
    • find . -name '*\.f' |xargs grep -r "string"
  • Number of files matching pattern 'pat*':
    • find . -maxdepth 1 -type f -name 'pat*' | wc -l
  • Number of empty files in present directory:
    • find . -maxdepth 1 -empty -type f | wc -l
  • Search only in non-binary files of all types:
  • find . -type f -print | xargs file | grep text | cut -f1 -d: | xargs grep -ir "some string"
  • Number of days in month $mm, year $yy:
    • dmax=$(cal ${mm} `date +${yy}` | egrep -v '[A-Za-z]' | wc -w);
  • rsync -a -e ssh storethis/ uname@ip:storehere/
  • LSF stuff:
    • Create group: bgadd -L 5 /groupname
    • Modify number of jobs in group: bgmod -L 1 /groupname
    • List group stats: bjgroup -s /groupname
    • Submit into group with bsub flag: bsub -g /groupname
    • Change queue: bswitch swell 29394929
    • Delete group: bgdel /groupname
    • Detailed job info: bjobs -l 9292929
  • ps/eps whitespace cropping: 
    • ps2eps -B -C < input.ps > output.ps
    • eps2eps Figure.eps Figure_Cropped.eps
    • ps2epsi input.eps output.eps
  • Simple PDF concatenation:
    • pdftk input1.pdf input2.pdf cat output output.pdf
  • Extract first word/number from each line of a file:
    • sed 's/ .*//' file
    • cat filename | cut -d " " -f1
  • Extract only integers from file:
    • grep -E '^[0-9]+$' in.txt > out.txt
  • Replace DOS ^M carriage returns with newline character:
    • tr '\015' '\n' < input_file > output_file
  • Bulk rename:
    • ls *txt|sed 's/\(.*\)\.txt/mv \1.txt  \1.m/'|sh
  • Find missing files from pairs (here, d+ -> single integer delimiter):
    • for fname in dep*.nc; do nums=`echo $fname | perl -wlne 'print $1 if /(\d+)/'`; if [ ! -f fpt.${nums}.nc ];then echo "fpt.${nums}.nc not found";fi; done
    • More complicated: for fname in dep.*.nc4; do nums=`echo $fname | perl -wlne 'print $1 if /(\d+\.\d+\.\d+.\d+)/'`; if [ ! -f fpt.${nums}.nc4 ];then echo "fpt.${nums}.nc4 not found";fi; done
  • Emacs hard-wrap adjustment
    • M-x auto-fill-mode
    • C-u 72 C-x f
  • FFMPEG syntax:
    • ffmpeg -f image2 -r 10 -i ./frame%d.png -b 600k ./atriv_hits.avi
  • FFMPEG splitting:
    • ffmpeg -i atriv_hits_12hr.avi -an -vcodec copy -ss 00:00:02 -t 00:10:08 atriv_hits_12hr_1970-1980.avi
  • Circumvent segfault when saving from MATLAB in nodisplay mode:
    • set(gcf,'Renderer', 'painters'); 
  • Default to white background in all MATLAB figures:
    • set(0,'defaultfigurecolor',[1 1 1])
  • Semi-magical debugging in MATLAB:
    • dbstop if error
      runme(someargument)
  • Use latexdiff seamlessly with git (assumes you've installed latexdiff):
    • Create ~/bin/git-latexdiff.sh with:
      • #!/bin/bash
        TMPDIR=$(mktemp -d /tmp/git-latexdiff.XXXXXX)
        latexdiff "$1" "$2" > $TMPDIR/diff.tex
        pdflatex -interaction nonstopmode -output-directory $TMPDIR $TMPDIR/diff.tex
        open $TMPDIR/diff.pdf
        rm -rf $TMPDIR

    • Let git run it:
      chmod a+x ~/bin/git-latexdiff.sh
    • Add the following to ~/.gitconfig:
      • [difftool.latex]
        cmd = ~/bin/git-latexdiff.sh "$LOCAL" "$REMOTE"
        [difftool]
        prompt = false
        [alias]
        ldiff = difftool -t latex

    • To run, do:
      • git ldiff somefile.tex
      • (compare with the standard method:)
         git difftool somefile.tex
      • Produces this nice PDF:
Picture