I have the following
"/x/x y/asfas/g.pdf " "fdfdf "/x/y/yy y/d.doc " "fdfdf etc.. I want to remove the spaces between the extension and the " ONLY
when I used tr to remove trailing space it removes the spaces in between that are more than 1 space (ex: yy y becomes yy y)
GNU sed approach:
sed 's/\([^.]*\.[^."[:space:]]*\)[[:space:]]\+"/\1"/' file \([^.]*\.[^."[:space:]]*\) - the 1st captured group containing hypothetical filename with extension
[[:space:]]\+" - ensures at least one space between extension and "
You could do:
sed 's/\(\.[^"[:blank:]]*\)[[:blank:]]*"/\1"/g' That is remove the sequences of blanks that follow . followed by a sequence of non-blank-nor-" characters and are followed by a "
To remove the blanks before every other double quote character on each line, you could do something like:
sed 's/[[:blank:]]*"/"&/g s/\(\([^"]*"\)\{3\}\)[[:blank:]]*"/\1"/g s/"\([^"]*"\)/\1/g' That is:
" before each <blanks>"<X>"<blanks>"<Y>"<blanks>" to <X>"<blanks>"<Y>""" to undo the insertions in 1Or in a more straightforward manner with awk:
awk -F \" -v OFS=\" ' { for (i = 2; i <= NF; i += 2 ) sub(/[[:blank:]]*$/, "", $i) print }' Note that tr wouldn't squeeze characters unless you use the -s option. More likely you forgot to quote a parameter expansion or command substitution that contained the output of tr.
In any case, tr can't be used for that task. It's just a transliteration tool. All it could do is translate/delete/squeeze all space characters. It cannot translate/delete/squeeze only some space characters.
Use GNU sed. Search for one or more spaces, followed by a double quote. Replace that with a double quote:
sed -i -e 's/ \+"/"/' file or like this, at your option:
sed -i -e 's/[[:space:]]\+"/"/' file 
"foo.bar" "baz" to "foo.bar""baz" though
"character?