Shell Programming and Scripting

I'm trying to use awk to extract certain text from a file. The file looks like this:

Code:

Start
1  10
2  20
3  30
 ...
 ...
 ...
<some word>

Now what I want to extract is the list of numbers (without the search text printed) that occur between "Start" and <some word>. The catch is that <some word> isn't always the same word!

I've gotten close with an awk script that looks like this:

Code:

awk '/Start/ {flag=1;next} /End/ {flag=0} flag {print}'  <filename>

This gives me exactly what I want, provided that <some word> happens to be "End" - problem is that it could be something else. The only sure thing is that it's some kind of letter. So I need to replace /End/ with some kind of variable that says to search between "Start" and the next time a letter A..Z is found.

I thought it should be fairly straightforward but I can't find a way (searched the boards). I'm not looking for anything too elaborate, so please don't spend too much time with complicated code! I have a workaround - just would like to see if there is a cleaner way to do it!

Try:

Code:

str="End"
awk '/Start/ {flag=1;next} $0~s {flag=0} flag {print}' s="$str" infile

Another one:

Code:

awk '/Start/{p=1;next}/End/{exit}p' file

Thanks both of you for the fast answers! Unfortunately, they didn't work for my case.

I'm not sure if I'm understanding the coding you are suggesting. Like I said in my initial post, the end of the numerical sequence doesn't occur necessarily with the word "End" - both of your code examples seem to be written as such, so I wasn't sure if I was supposed to replace that with something else.

As I said, the difficulty is that I don't know what is going to end the numerical list. All I know is that it's going to be some kind of word. is there a way to replace the search string /End/ within awk to say that as soon as it encounters any letter from A .. Z, stop the command?

OK, then I think you need this:

Code:

awk '/Start/ {flag=1;next} /[a-zA-Z]/ {flag=0} flag {print}'  <filename>

Weird. That looks perfect - it should do what I want. But when I try that, I only get the one line that has the word Start in it.

I'm using AIX and ksh, if that makes any difference.

I tried doing it backwards by saying find anything BUT a number:

Code:

awk '/Start/ {flag=1;next} /^[0-9]/ {flag=0} flag {print}'  <filename>

... but that didn't work either.


EDIT: I realized that the above code means it will stop when it encounters something that is not a number, but at the BEGINNING of the line. Is that true?

I'd like it to stop extracting as soon as it sees a letter even if it doesn't occur at the beginning of the line.

Code:

awk '/Start/ {flag=1;next} /[A-Za-z]/ {flag=0} flag {print}'  <filename>