Shell Programming and Scripting

I've been working on this all night and finally have to ask for help... and not just from my coffee pot.

I need to replace a line of text only when it is proceeded by a line containing only the letter "H"

Input:

Code:

H      -2.204711   -0.922090   -0.024814

P
6-311+G(d)
****
C
6-311+G(d)
****
H
6-311+G(d)
****

Desired Output:

Code:

H      -2.204711   -0.922090   -0.024814

P
6-311+G(d)
****
C
6-311+G(d)
****
H
6-311++G(d,p)
****

I've tried many variations of perl and sed code that I have found here and elsewhere. As a result I've managed to do the opposite of what I wanted or to make a global substitution in dozens of different ways, but never have I managed the desired substitution.

My most recent attempts have been of this general nature:

Code:

perl -pe '$i=1 if (m+H+);if(m+FIND+) {if($i=1){s/FIND/REPLACE/} else {$i=0}}' input > output

I'm open to any options and am only using perl because it's what I'm most familiar with.

While the find and replace expressions I've been using for the above example may be incorrect I know that there is also an issue with the substitution code because I have used it on a simplified example containing only friendly characters and have encountered the same issues.

sed can do the job.

Code:

sed 'N;/^H/s/(d)/(d,p)/g' input > output

Thank you so much!

Me thinks it's time to sit down and learn some sed and awk.

Something not working at all is one thing, but something inexplicably working sometimes is an even great frustration.

The code posted by danmero works great for the code example I had given, but performs no substitution when given the following as input:

Code:

H	 3.228216   -1.558532	-0.403837

P
6-311+G(d)
****
O N
6-311+G(d)
****
C
6-311+G(d)
****
H
6-311+G(d)
****

I use:

Code:

sed 'N;/^H/s/+G(d)/++G(d,p)/g' in > out

and no change is made in the output. Anyone have any ideas? I'm really confused because it seems to me that the conditions for the substitution don't really change at all, and the condition appears in this input file exactly as it does in the input which work with the sed code.

Code:

perl  -00 -pe's/(\n?H\n.*?)(\Q+G(d)\E)/$1++G(d,p)/g' file

With awk (nawk or /usr/xpg4/bin/awk on Solaris):

Code:

awk 'f{sub(/\+G\(d\)/,"++G(d,p)");f=0}1;/^H$/{f++}' file

or sed:

Code:

sed '/^H$/{n;s/+G(d)/++G(d,p)/;}' file

Strange, my solution work for your second example. Next time try to provide some informations about your work environment.
Why you need a second plus ?

The sed code has worked properly on all the inputs I've tried so far.
Thank you very much for your time.

Now if only I could figure out what the critical difference is...