[SOLVED] How to extract the first word following a string

Feynman · 08-23-2010, 12:23 PM

crts--I just tried that and wrote in test2:
105
46

You can see where those numbers came from if you look back at the line. Interesting result. Thank you all for staying with me. I really appreciate it.

crts · 08-23-2010, 12:26 PM

Quote:

Originally Posted by Feynman

crts--I just tried that and wrote in test2:
105
46

You can see where those numbers came from if you look back at the line. Interesting result. Thank you all for staying with me. I really appreciate it.

Try

Code:

sed -nr "s/.*${pattern}[      ]*([^     ]+).*/\1/p" /opt/gamess/tests/exam12.inp > test2

The problem is that \w will get a whole word. The '.' is considered a word boundary and therefor not part of the result. Since you want 46.xxx as result \w is not suitable. This was not obvious from your first post. That is why you should always provide sample data as accurate as possible.

P.S.: If you can be certain that there are no tabs in your file you can omit the tab characters and use only spaces inside [] instead of both.

Feynman · 08-23-2010, 12:40 PM

crts--when I tried your script test2 read
105
46

You can see where these numbers came from if you look at the line I am extracting from. Weird result.
Thanks for staying with me.

Feynman · 08-23-2010, 12:42 PM

sorry ignore that last post. I did not notice the "next page feature"

Feynman · 08-23-2010, 12:45 PM

Sorry, I will try to be clearer. Now I got
105.14088
46.45648

How do I just get the first number?

crts · 08-23-2010, 12:57 PM

Quote:

Originally Posted by Feynman

Sorry, I will try to be clearer. Now I got
105.14088
46.45648

How do I just get the first number?

Based on your sample data

Code:

I tried typing the input by the forum seems to automatically take out extra spaces (obviously ctrs found a way around that). Here is what the particular line would look like:
![3 spaces]Heat OF FORMATION[9 spaces]105.14088[3 spaces]93.45997[3 spaces]46.89387

you already got the first number. The number 46.45648 is from another line. It is not 46.89387 form the same line. Notice, that they are not equal. If you do not want to process anymore lines after the first match then try

Code:

sed -nr "/${pattern}/ {s/.*${pattern}[      ]*([^     ]+).*/\1/p;q}" /opt/gamess/tests/exam12.inp > test2

Maybe it is best if you post your data file as attachment and tell us what data you want extracted.

Feynman · 08-23-2010, 01:06 PM

That worked great! Thank you very much! I wanted to attach my data but could not figure out how. Next time!

crts · 08-23-2010, 01:28 PM

Quote:

Originally Posted by Feynman

That worked great! Thank you very much! I wanted to attach my data but could not figure out how. Next time!

When you reply (not quick reply) look out for the "paper-clip" symbol. Keep in mind that attaching is "extension" sensitive. You will have to rename the file like

filename.txt

Otherwise you won't be able to upload it.

konsolebox · 08-24-2010, 03:00 AM

Quote:

Originally Posted by crts

@konsolebox: I am also not quite sure what you mean. Since sed uses the -r switch for extended regex there should be no need to escape the braces.

When has sed implement -r of super sed... I never thought about it.

Tinkster · 08-24-2010, 12:27 PM

Quote:

Originally Posted by konsolebox

When has sed implement -r of super sed... I never thought about it.

Not sure when exactly it got introduced, but I have
an inkling it was around the version jump from 3.xx -> 4.xx

I've certainly been using it for a long time.

Cheers,
Tink