UNIX for Dummies Questions & Answers

How do I search for a string within a file and return the names of the file that contain the string? I would like to search directories and sub-directories.

Code:

find /top/directory -type f  -exec grep -l 'string I want to find' {} \;

#!/usr/bin/ksh
#
# usage: stringfinder.ksh string_to_find filesystem_to_search [eg:stringfinder.ksh test /opt/local/bin]


if [[ -z "$1" ]]; then
print -n "Which string/word/number are you searching for? : "
read STR
else
STR="$1"
fi

print "Will search for ${STR}"
if [[ -z "$2" ]]; then
print -n "Which filesystem would you like to search? : "
read FSYS
else
FSYS="$2"
fi

print "Looking for ${STR} in ${FSYS}"

print "FILES WITH ONE OR MORE INSTANCE OF ${STR}:"

for i in `find ${FSYS} -print`
do
export DIR=${i}
CT=`grep -i ${STR} ${i}|wc -l`
if [ ${CT} -ge 1 ];then print " ${DIR} has ${CT} instances";fi
done

print "Done searching for ${STR} in ${FSYS}"

unset STR
unset FSYS
unset DIR
unset CT

exit 0

searchme.sh: line 13: print: command not found

I get the error as mentioned above.

print was not a standard command in the original Bourne shell, and doesn't exist in all dialects (tho I would have thought it always existed in ksh!?). You can use echo instead.

The -n is supposed to suppress a newline after the printed message; your echo might use a different convention, such as wanting a \c at the end of string rather than the -n option. Check the manual page.

Jim's solution is more elegant, though. (And the scripter should use grep -c rather than grep | wc -l)