Hi

I need sed or awk script which basically takes the file containing the input as below

Original file
apple
bat
elephant
cat
dog
elephant
fox
goat
elephant
cat
dog
fox
elephant
fox

output should be like
apple
bat
elephant
cat
giraffe
dog
fox
goat
elephant
cat
giraffe
dog
fox
fox

Basically, First I want to search for line containing "cat" and from there on search for "elephant" and delete the line containing elephant and add line "girrafe" after cat.

Request for speedy answer if possible

Sounds like homework.

put in some effort....show us what you have got.

Okay ..
I tried to solve this by breaking the problem into two steps
1 st step - Add Giraffe after cat blindly
2nd step - search between cat and ) (ofcourse its non existing here) and delete elephant.
I was able to get that, but I am missing the

parent bash script contains..says parent.sh
>>>
sed '/cat/a\Giraffe' hyd1 > hyd2
sedfile3 hyd2 > hyd3
<<<

sedfile3 contains ..(sedfile3)
>>>
sed -n '
/cat/,/)/ {
s/Ele.*//
s/[ ^I]*$//
/^$/ d
p
}
' $1
<<<

And If I run ./parent.sh I get output as
cat
giraffe
dog
fox
goat
elephant
cat
giraffe
dog
fox
fox

not as intended .. I am missing the portion before "cat"

Some typo in my previous message

1 st step - Add Giraffe after cat blindly
2nd step - search between cat and ) (ofcourse its non existing here) and delete elephant.
I was able to get that, but I am missing something.


hyd1 file contains. ..
apple
bat
elephant
cat
dog
elephant
fox
goat
elephant
cat
dog
fox
elephant
fox

My parent bash script contains (name parent.sh)
>>>
sed '/cat/a\Giraffe' hyd1 > hyd2
sedfile3 hyd2 > hyd3
<<<



sedfile3 contains ..(sedfile3)
>>>
sed -n '
/cat/,/)/ {
s/Ele.*//
s/[ ^I]*$//
/^$/ d
p
}
' $1
<<<

And If I run ./parent.sh I get output as
cat
giraffe
dog
fox
goat
elephant
cat
giraffe
dog
fox
fox

not as intended .. I am missing the portion before "cat"

I have gone through the backup questions, but unable to find on how to search the second pattern elephant ( and delete), between pattern cat (and say fox)

The US is not awake yet, and Europe is already at work. You may have to be patient.

I find this a bit hard to follow.....two things:

1. You use "sed -n" which suppresses printing until a "p" is found. Is that why you are missing all before "cat".

2. To search the second pattern, you may need this general form:
sed -e '<some code>' -e '<some more code>' oldfile > newfile

I don't understand this.


Finally, what references are you using for SED?

i don't really understand your output, it seems like you only want to delete the first instance of "elephant" after cat, right?

Let me explain,
I wanted is to search for "elephant" between patterns "cat" and "fox" and delete the line containing "elephant"


If original file contains:
apple
bat
elephant --- Ignore,this elephant as this is notwithin the patterns "cat" and "fox"
cat
dog
elephant -- I have to remove this line as is it is between "cat" and "fox"
fox
goat
elephant - Ignore
cat
dog
fox
elephant - Delete
fox


Output should be

apple
bat
elephant
cat
dog
fox
goat
elephant
cat
dog
fox
fox

HTH in better understanding of my problem.

Okay guys, I was able to solve the problem by following the below mentioned approach

parent.sh contains
>>>.
sed '/cat/a\Giraffe' hyd1 > hyd2
sedfile3 hyd2 > hyd3
<<<<

sedfile3 contains
>>>
sed '
/cat/,/fox/ {
s/ele.*//
s/[ ^I]*$//
/^$/ d
}
' $1
<<<<<

It is giving me the desired output.
But do let me know, if there is any easier approach.
This was a really learning experience on sed.

This almost works:
The problem is the last instance where you have:
The problem is that the SED addressing range goes from "cat" to the next instance of "fox"---thus it never sees the last instance of "elephant".

The real issue in this kind of problem is to unambiguously define the logic to be used. SED addressing goes from <address1> to the next instance of <address2>. What you need is logic that goes from <address1> to the last instance of <address2> (before finding <address1> again). But the latter logic might not work with some other pattern in the target file.......

I'll post back if I think of something....

Is there a way, where in I want to search within the pattern (say here between "cat" and "fox" , modify "elephant" to "dragon", but only in the first range.
For eg:
If input is like
cat
dog
elephant
fox
cat
dog
elephant
fox

I want output like
cat
dog
dragon << only elephant in first range block of cat and fox is modified"
fox
cat
dog
elephant << not the second one in the range of "cat and fox"
fox

I want to find and replace only first occurance in a block.
How to do that using sed.
Say,

Word1
a
b
c
Word2
Word1
a
b
c
Word2

is there in my input file.

My script
sed '/Word1/,/Word2/{s/a/d}' file

will give the output replacing all a's to d's between this word range.
But I want to restrict in only first block.
My output should be like

Word1
d
b
c
Word2
Word1
a
b
c
Word2

Any suggestions ..

Looks like homework to me. Why don't you try eg. explicitly specifying lines 2-4 for replacement. Would it work in your case?