How can I delete all of the lines in a file which contain more than a given number of letters? E.g.
bear rabbit tree elephant If I restrict it to words of 5 letters or less, the output would be:
bear tree 
$ awk 'length<=5' input.txt bear tree The following would do the trick:
sed -i '/^.\{5,\}$/d' FILE What that means is this:
Delete (/ [...] /d) in-place (-i switch) all lines matching the following pattern:
^).) repeated 5 or more times (\{5,\})$)from the file named FILE.
grep -v '......' myfile.txt will deliver lines five characters or less.
It does this by "selecting" lines containing six characters or more, then reversing the action with -v, to only print out those that don't match.
grep -Ev '.{6,}' is more general and golfs identically in this case. It is however faster to type 6 dots than .{6,}."The file contains various foreign characters, each of these should count as one letter." Assuming your input data is in UTF8, this bash filter script should do it.
#!/bin/bash function px { local a="$@" local i=0 while [ $i -lt ${#a} ] do printf \\x${a:$i:2} i=$(($i+2)) done } (iconv -f UTF8 -t UTF16 | od -x | cut -b 9- | xargs -n 1) | if read utf16header then px $utf16header cnt=0 out='' while read line do cnt=$(($cnt+1)) if [ "$line" == "000a" ] then if [[ $cnt -le 5+1 ]] ; then out=$out$line px $out fi cnt=0 out='' else out=$out$line fi done fi | iconv -f UTF16 -t UTF8 