Shell Programming and Scripting

Hi,
I have a large file with a repeating pattern in it. Now i want the file split into the block of patterns with a specified no. of lines in each file.
i.e. The file is like

1...
2...
2...
3...
1...
2...
3...
1...
2...
2...
2...
2...
2...
3...

where 1 is the start of the block and 3 is the end of the block.
Now i want the file to be split up in blocks. But each file can have more than one block, but specified no. of lines or less.

What i have done is i split up the file using

Code:

awk '/^1/{close("file"f);f++}{print $0 > "file"f}'  testfile

But this gives one file for each block. I needed to concat the files to get a file with some specified no. of lines, say 10000.

Please give me an efficient way to do this. I feel we can do it in the awk command only.
Thanks,
CSS

Code:

awk 'NR%10000==1{close("file"f);f++}{print $0 > "file"f}'  testfile

Regards

But that is writing 10000 line per file. But not patternwise.

Maybe this will help:

Code:

awk  '/^1/{f=1} f{ print $0 > "file_"n ; c++} c==x+1{ n++; c=1; close("file_"n)} /^3/{f=0}'  x=10000 c=1 n=1  filename

Change x to whatever number of lines per file you need to have. Whereas c and n are just counters, that don't need to be changed.

Even this is giving files with 10000 lines but i want to split in patterns.

Thanks Rubin,
But i had to make some minor changes and this worked.
awk '/^1/{f=1} f{ print $0 > "file_"n ; c++} c>10000 && /^3/ { n++; c=1; close("file_"n) }' c=1 n=1 testfile