Change path using awk or sed

Question

Hello i want to change symlink path to real path in txt file. First I find word "SF:" and change path. Now i use this script but it is slow and not effective. I think it can be changed by awk or sed command. Thank you in advance.

#!/bin/bash
FILENAME="new1.info"
    echo "" > $FILENAME
while read LINE
do
if [ "" != "$(echo $LINE | grep -e "^SF:")" ]
then
    echo "SF:""$(realpath $(echo $LINE | cut -d":" -f2))" >> $FILENAME
else
    echo $LINE >> $FILENAME
fi
done < total.info

I have big txt file, so I want to find "SF:" and change line from something like this:

SF:/home/ects/svn/moduleTests/ctest/tests/moduletests/base/tmp/src/base64.cpp

To this:

SF:/home/ects/svn/moduleTests/ctest/src/base/base64.cpp

Hello @george123. Welcome to SO. Have you made some reasearches yet? Could you please show us what you have already tried? Also consider reading this: https://stackoverflow.com/help/how-to-ask — Maël Pedretti, Sep 19 '18 at 09:34
Try `sed -i 's/^SF://'` (or, `sed -i '' 's/^SF://'` if you are on Mac) if all you want is to remove `SF:` at the start of each line — Wiktor Stribiżew, Sep 19 '18 at 09:34
make a small example, what you have, and what do you want to get. we don't know what text in your `$LINE` — Kent, Sep 19 '18 at 09:53
I went looking for a solution for exactly the same problem, and I do mean exactly. I assume you're doing this because the SonarTS plugin in SonarQube doesn't handle symlinks? — David M. Karr, Dec 15 '20 at 16:46
No, it was related to the LCOV code coverage tool, but this tool doesn't handle symlinks too. — jorginho97, Dec 16 '20 at 17:21

glenn jackman · Accepted Answer · 2018-09-19T10:45:09.280

In bash, I'd write

#!/bin/bash
while IFS= read -r line; do
    if [[ "$line" == "SF:"* ]]; then
        line="SF:$(realpath "${line#SF:}")"
    fi
    echo "$line"
done < total.info > new1.info

Things to note:

quote your variables, every time.
don't use ALLCAPS varnames, it will bite you eventually
note the output redirection has been moved to the outside of the while loop, just like the input redirection
IFS= read -r line is the way to read lines from the file exactly.

I don't know if this will be any faster: bash can be quite slow, particular for while read loops over big files. You could try another language:

perl -MCwd=abs_path -pe 's/^SF:\K(.*)/ abs_path($1) /e' total.info > new1.info

@george123, for your future questions, please read https://stackoverflow.com/help/someone-answers — glenn jackman, Sep 19 '18 at 14:20

Change path using awk or sed

1 Answers1

Linked