negate search help

Shell Programming and Scripting
1

Hi, I've tried a lot of negate codes in this forum, but they do not perform what I intended. Please help.

inputfile:

Paragraph1 contents: die1, die2, die3, pr_name1, pr_name2
pr_name3, pr_name4

Paragraph2 more contents: die1, die2, die3, pr_name1, pr_name2
pr_name3, pr_name4

outputfile:

pr_name1
pr_name2
pr_name3
pr_name4

pr_name1
pr_name2
pr_name3
pr_name4

The 2 codes that I've found so far are:
sed -ne '/Paragraph1/s/Paragraph1 contents: die1, die2, die3, pr/pr/p' inputfile > outputfile
This only give me pr_name1, pr_name2. I want to modify this code sed -ne '/Paragraph./s/Paragraph.(anything in between)pr/pr/p' inputfile > outputfile

awk '/Paragraph1/{print $NF}' inputfile > outputfile
This only give me pr_name2.

thanks.

2 
$ cat infile 
Paragraph1 contents: die1, die2, die3, pr_name1, pr_name2
pr_name3, pr_name4

Paragraph2 contents: die1, die2, die3, pr_name1, pr_name2
pr_name3, pr_name4

$ perl -00>outfile  -lne's/, /\n/g;print/((?:.*?\n*){4})$/' infile 
$ cat outfile 
pr_name1
pr_name2
pr_name3
pr_name4

pr_name1
pr_name2
pr_name3
pr_name4
3

try somethig like this

sed -e 's/, /\
/g' -e '/^pr/!d' infile
4

Thanks, Radoulov for the perl script. I'm using .bashrc and perl is much more complicated. I think your code is looking at the 4th field after the last comma. My report varies so it's not always the 4th field, but I forgot to mention that :slight_smile:

Vidyadhar85, your codes works perfectly. Can you please explain how it works? I tried to combine the 2 lines together (sed -e 's/, /\/g' -e '/^pr/!d' infile) and of course it didn't work. hehe. I only thing that I understand is the first line look for the last comma using the -e option, but I dont understand the /\ part. Then the next line negate delete all words that don't start with ^pr. Please explain more in details so I can leverage your code in the future. Thanks.

5

basically it enteres a new line after each ", "
so i am searching for ", " and then inserting newline by hitting an enter after search thats why it went to next line so it won't work in single line
regards,
vidyadhar

6

Vidyadhar85, Thanks for the explaination. Leverage from your codes, I have modified it for anyone who might find this useful. I used:

sed -e 's/, /\n/g' -e '/pr/!d' -e 's/contents: /\n/g' input > temp
sed -e '/pr/!d' temp > output