Scripting - replace a line <SOLVED>

[dcung]

Hi,

I have 100 files in a directory.
I'd like to replace one line in each file with 2 new lines.

word1-word2-word3

with

word8-word9
word1-word2-word3 /word4/word5/word6/word7


If it was a line, my rusty memory remember that I should create a script that goes something like

for $i in *
do s/word1-word2-word3/word1-word2-word3 \/word4\/word5\/word6\/word7/

I don't mind doing it in two stages, i.e. substitute a line and later insert a line.

BTW - I use XenPup75 or Fatdog64.

Calling scripting gurus to help please...

Thank you.

[musher0]

Hello dcung.

Your choice of approaches!
https://www.ask.com/web?l=dir&qsrc=0&o= ... +new+lines

IHTH

[fredx181]

Hello dcung,

This should do from terminal in that directory:

Code: Select all

for i in $(ls -A)
do
sed -i "s|word1-word2-word3|word8-word9\nword1-word2-word3 /word4/word5/word6/word7|g" "$i"
done

Or set "word1-word2-word3" as variable first:

Code: Select all

VAR="word1-word2-word3"
for i in $(ls -A)
do
sed -i "s|$VAR|word8-word9\n$VAR /word4/word5/word6/word7|g" "$i"
done

Note that this will change your files because of the "-i" switch of sed.
Better first test with the -e switch on only a few files in a directory (and you'll just see the output in terminal, no files modified).

Code: Select all

for i in $(ls -A)
do
sed -e "s|word1-word2-word3|word8-word9\nword1-word2-word3 /word4/word5/word6/word7|g" "$i"
done

Fred

[dcung]

Or set "word1-word2-word3" as variable first:

Code: Select all

VAR="word1-word2-word3"
for i in $(ls -A)
do
sed -i "s|$VAR|word8-word9\n$VAR /word4/word5/word6/word7|g" "$i"
done

I used this one. Works like a charm!
Did it in a second, what would have taken me at least 3 days to do manually...
Thanks Fred.

[musher0]

Hi dcung.

May I suggest: (please read the notes and comments)

Code: Select all

#!/bin/bash
# /root/my-applications/bin/replace1linewith2.sh
#### # set -xe
B="word8-word9"
A="word1-word2-word3" 
C="/word4/word5/word6/word7"

TxT=nonsense # or whatever the file name
rm -f mod-$TxT* # Clean previous tries.

N=1
cd /root/my-documents # or wherever the documnents are.
ls -1 $TxT* > texts.lst
while read text;do
	echo . > mod-$text-$N # Adding empty line for readability
	while read line;do
		if [ "$line" = "$A" ];then
			echo -e "$B\n$A $C" >> mod-$text-$N 
		else
			echo $line >> mod-$text-$N
		fi		 
	done < $text	
done < texts.lst
# N="`expr $N + 1`" # Could be used for versioning.
# set +xe
# more /root/my-documents/mod-* # Optional viewing of results
exit

~~~~~~~~~~~~~~~~~~~~~~~
NOTES
-- for dcung, PuppyLinux Forum, Sat. Aug. 18, 2018.
-- Make sure the original text ends with an empty line. 
Otherwise, any script like this will skip the last line of text.
			
In this test, the following three files were used as starting points:
-- nonsense3.txt
folbolo folbolo
blobloblo
toto
word1-word2-word3
strum strum strum
bebop hiphop

-- nonsense2.txt
folbolo folbolo
blobloblo
tah tah
word1-word2-word3
strum strum strum
bebop hiphop

-- nonsense.txt
falbala falbala
blablabla
toto
word1-word2-word3
strum strum strum
bebop hiphop

[~/my-documents]>more mod-non*
.
folbolo folbolo
blobloblo
toto
word8-word9
word1-word2-word3 /word4/word5/word6/word7
strum strum strum
bebop hiphop
.
folbolo folbolo
blobloblo
tah tah
word8-word9
word1-word2-word3 /word4/word5/word6/word7
strum strum strum
bebop hiphop
.
falbala falbala
blablabla
toto
word8-word9
word1-word2-word3 /word4/word5/word6/word7
strum strum strum
bebop hiphop

IHTH