[SOLVED] Bash to read text file and replace a line

farenheitcx · 11-19-2011, 12:06 PM

I need a bash script to find and replace a line with the text "//status_id" to "status_id" and then save the modification into the same file.

I searched in the forum some options with "awk" and "sed" to delete the first two characters but i dont know how exactly to do that.

This is the content of the file to find and replace the text.

Code:

// THIS LINES HAVE A SIMPLE TEXT
// THIS LINES HAVE A SIMPLE TEXT
// THIS LINES HAVE A SIMPLE TEXT
// THIS LINES HAVE A SIMPLE TEXT
// THIS LINES HAVE A SIMPLE TEXT
// THIS LINES HAVE A SIMPLE TEXT

// PERFORMANCE PARAMETERS
cl_graph "32"
mp_inco "3.5"
cl_ficas "22"

// CHECK STATUS
//status_id


cl_graph "32"
mp_inco "3.5"
cl_ficas "22"
cl_graph "32"
mp_inco "3.5"
cl_ficas "22"

sycamorex · 11-19-2011, 12:14 PM

Well you could simply replace //status_id with status_id:

Code:

sed 's/\/\/status_id/status_id/' file

or without escaping forward slashes:

Code:

sed 's@//status_id@status_id@' file

farenheitcx · 11-19-2011, 12:34 PM

Many thanks for your quickly reply. I need some help to perform this script for multiple search and replacing words. I try with this but something is wrong..

Code:

#!/bin/bash
ftxt="//status_id cl_ficas" #find words to replace
rtxt="status_id //cl_ficas" #replace words with this

for u in $ftxt 
do
	for r in $rtxt
	do
		sed 's/\/\/'"{$u}"'/'"{$r}"'/' file
	done
done

sycamorex · 11-19-2011, 12:46 PM

edit: disregard. I didn't read it properly

For variable substitution you need to use double quotes ("). Try and see what happens.

sycamorex · 11-19-2011, 12:51 PM

edit: disregard the commands below. It obviously is wrong. the forward slashes are already included in the variables.

I'm not sure if you haven't made it overcomplicated. Try:

Code:

sed "s/\/\/$u/$r/g" file

Once you're happy with the results, add the -i flag to make the changes in the file.

Code:

sed -i "s/\/\/$u/$r/g" file

farenheitcx · 11-19-2011, 01:01 PM

Quote:

Originally Posted by sycamorex

I'm not sure if you haven't made it overcomplicated. Try:

Code:

sed "s/\/\/$u/$r/g" file

Once you're happy with the results, add the -i flag to make the changes in the file.

Code:

sed -i "s/\/\/$u/$r/g" file

I will be testing

farenheitcx · 11-19-2011, 01:15 PM

Nothing happend. I receive this error

Quote:

sed: -e expression #1, char 19: unknown option to `s'

Code:

#!/bin/bash -x
ftxt="//status_id cl_ficas" #find words to replace
rtxt="status_id //cl_ficas" #replace words with this

for u in $ftxt;
do
        for r in $rtxt;
        do
                sed -i "s/\/\/$u/$r/g" file
        done
done

sycamorex · 11-19-2011, 01:17 PM

Sorry, I've been talking to someone and writing on LQ. Hence, rather chaotic and incorrect replies. My apologies. Can't concentrate on 2 things at the same time.

Now, do you need to use loops and variables? Your whole script could be replaced with 1 line:

Code:

sed "s@//status_id@status_id@;s@cl_ficas@//cl_ficas@" file1

farenheitcx · 11-19-2011, 01:27 PM

Okay, I appreciate your help.
sycamorex good job, thank you for solving my problem, thank you very much!

Solved