Recursively search and extract substring from files in current dir using shell scripting in Linux?

Question

I want to recursively search and extract in all files inside my current directory, for any string that starts with " B-0 " followed by any number of digits .

If a match is found, i want to extract it. But at the same time, i want to extract unique strings . For example - search might find B-05255 in 2 files. So i want to extract B-05255 only once (unique).

Once the search is done, i want to allocate all strings to one final string comma separated as :

FINAL_STRING = B-05255,B-05256,B-04152

Waiting for suggestions?

Thanks.

Answer 1

You can use the option -r (recursive) :

GREP_OPTIONS="" /bin/grep -ohr '\bB-0[0-9]*' . | tr $'\n' ','

This iterates recursively trough the currently directory and searches for the pattern in every file. The option -o extracts only the matched strings (not the whole line). Note that I'm using the escape sequence \\b which stands for word boundary (Because you said: "... that starts with ...")

However grep will print each result on a separate line. I'm piping the results to tr to replace newlines by commas.

Answer 2

You can use this grep , sort , tr combination:

FINAL_STRING=$(grep -rIhEo '\bB-0[[:digit:]]*' . | sort -u | tr '\n' ',')
echo "${FINAL_STRING%,*}"

grep options used are:

• r - recursive
• I - ignore binary files
• h - omit filename in output
• E - extended regex
• o - only print matched output