Pattern Search Fixed width Position

[lnair]

Record length is 820 Characters fixed width position with 2 million rows ///


112223333sugar458696963
125632233water4589sugar
125632233people63336666

Output would be something like this


Output:

File 1
-------
112223333sugar4586969633

File 2
-------
125632233water45869sugar
125632233people633366666

[harry d brown jr]

If you use the command line it will take TWO passes through the data:

# grep "^\([0-9]\{9\}\)sugar\(.*$\)" testfile1 > outfile1
# grep -v -f outfile1 testfile1
125632233water4589sugar
125632233people63336666
# grep -v -f outfile1 testfile1 > outfile2
# cat outfile1
112223333sugar458696963
# cat outfile2
125632233water4589sugar
125632233people63336666
# cat testfile1
112223333sugar458696963
125632233water4589sugar
125632233people63336666
# grep "^\([0-9]\{9\}\)sugar\(.*$\)" testfile1 > outfile1
# grep -v -f outfile1 testfile1 > outfile2
# cat outfile1
112223333sugar458696963
# cat outfile2
125632233water4589sugar
125632233people63336666
#

Because it requires TWO passes and a LOT of OVERHEAD, I suggest you write a perl script to do it.


live free or die
harry

[harry d brown jr]

#!/usr/bin/perl
#
open(ORIGfileptr,"< ORIGFILE");
open(OUToneFileptr,"> file1");
open(OUTtwoFileptr,"> file2");
while () {
if ( m/^([0-9]{9})sugar(.*$)/ ) {
printf OUToneFileptr $_;
} else {
printf OUTtwoFileptr $_;
}
}
close(ORIGfileptr);
close(OUToneFileptr);
close(OUTtwoFileptr);


change ORIGFILE (on the open line (the one in quotes)) to the real file name.


live free or die
harry

[Jeroen Peereboom]

I do not understand your question. Maybe you can clarify a little bit.

* What is the criterion to put a line in file 1 or file 2? The first number (9 digits)?
* Only 2 output files?

JP.

[lnair]

Thanks for the Suggestion;

However I was looking for an slightly different answer; to be precise, Pattern "Sugar" could be anything (non empty characters);

So basically the comparison would be something like this For instance(from position 10-14 non-empty characters spooled to 1 file and rest empty characters spooled to 2 file )


112223333sugar458696963
125632233water4589sugar
125632233people63336666

[harry d brown jr]

#!/usr/bin/perl
#
open(ORIGfileptr,"< ORIGFILE");
open(OUToneFileptr,"> file1");
open(OUTtwoFileptr,"> file2");
while () {
if ( m/^([0-9]{9})([A-Za-z]{5})(.*$)/ ) {
printf OUToneFileptr $_;
} else {
printf OUTtwoFileptr $_;
}
}
close(ORIGfileptr);
close(OUToneFileptr);
close(OUTtwoFileptr);

The change says that any ALPHA string five characters long in positions 10-14, go to "file1", all others go to "file2".

TO SHOW SPACES in your examples, use the "+" sign.

live free or die
harry