CODEWITHSUNDEEP
sed
user1074593
user1074593

Reputation: 630

Search for string in sed after ocurrence of another string

The input file I have looks like this

Topic1
Some text
Topic2
Some text
Topic3
Some text

I need to go to Topic2 and then replace "Some text" with "Actual input"

The following script works

/Topic2/,/Topic/ {
/Some text/c
Actual input
}

Is there a simpler way to do this. I was hoping the following will work but it does not.

/Topic2/ {
/Some text/c
Actual input
}

Update : Should have made it clearer. Am only looking for sed based solutions.

Upvotes: 0

Views: 76

Answers (3)

potong
potong

Reputation: 58473

This might work for you (GNU sed):

sed -e '/Topic2/,/Topic/{//!d;/Topic2/a\bla\nbla\nbla' -e '}' file

or:

sed -e '/Topic2/,/Topic/{//!d;/Topic2/r fileb' -e '}' filea

or using Bash:

sed $'/Topic2/,/Topic/{//!d;/Topic2/a\\bla\\nbla\\nbla\n}' file

EDIT:

sed $'/Topic2/{:a;n;/Some text/!ba;c\\Actual input\n}' file

Upvotes: 1

Jotne
Jotne

Reputation: 41460

Here is a gnu awk solution:

awk -v RS="Topic[0-9]*" 'f=="Topic2" {$0="\nActual input\n"} NR>1{print f,$0} {f=RT}' file
Topic1
Some text

Topic2
Actual input

Topic3
Some text

Upvotes: 0

Kent
Kent

Reputation: 195179

how about with awk?

awk 'p&&/Topic/{p=0}/Topic2/{p=1}p{gsub(/Some text/,"foo\nbar\nbla\nline")}7' file

above I used a multiple line replacement as example, the output with your input would be:

Topic1
Some text
Topic2
foo
bar
bla
line
Topic3

you can pass the replacement string to awk as shell variable, so that the awk one-liner would be flexible. like:

kent$  r="a
b
c"

kent$  awk -v rep="$r" 'p&&/Topic/{p=0}/Topic2/{p=1}p{gsub(/Some text/,rep)}7' f  
Topic1
Some text
Topic2
a
b
c
Topic3
Some text

Upvotes: 1

Related Questions