[SOLVED] Pull sections of text from a file and create a new file?

soupmagnet · 12-26-2012, 10:34 PM

I'm not quite sure how to explain this, but here it goes anyway.

I have rather large text files that are formatted like so:

Code:

FIELD 1
1 This is the first string
2 This is the second string
3 This is the third string
4 And so on
5 And so forth

FIELD 2
1 And so on
2 And so forth
3 Each field having an indeterminate number of strings
4 And there may be an indeterminate number of fields

FIELD 3
1 The only usable patterns (that I see) are
2 The titles of each field are numbered (anywhere between 1 and 150)
3 Each string in a field is numbered (also anywhere between 1 and 150)
4 And there are empty lines between each field

I need to create a new file for every single field in the original file. Each file needs to have the information and preferably the title of each individual field. And I would need to script this because I have several hundred of these "original" files.

Is this even possible?

anon237 · 12-27-2012, 02:43 AM

Is this what you are looking for:

Code:

awk 'BEGIN { RS="" ; FS="\n"} { outfile = $1 ; for ( i=1;i<=NF;i++) print $i > outfile }' infile

David the H. · 12-27-2012, 04:35 PM

I've been waiting for someone to ask a question like this.

Code:

csplit input.txt '/^FIELD/' '{*}'

csplit is one of the lesser-known coreutils. It has several options for formatting the output file names and such too.

The only minor drawback in this case is that the empty lines between the fields are not removed, but still show up at the end of the files. But that could be corrected fairly easily if needed with sed, or perhaps grep -v.

Edit: You can remove the trailing blank lines with this after you're done:

Code:

for fname in xx* ; do
	printf '%s\n' '/^$/d' 'w' | ed -s "$fname"
done

soupmagnet · 12-27-2012, 04:46 PM

Quote:

Originally Posted by druuna

Is this what you are looking for:

Code:

awk 'BEGIN { RS="" ; FS="\n"} { outfile = $1 ; for ( i=1;i<=NF;i++) print $i > outfile }' infile

Works perfectly, thank you.

Now to figure out what it all means. It looks like I'll have to study up on 'awk' some more.

Quote:

Originally Posted by David the H.

I've been waiting for someone to ask a question like this.

Code:

csplit input.txt '/^FIELD/' '{*}'

csplit is one of the lesser-known coreutils. It has several options for formatting the output file names and such too.

The only minor drawback in this case is that the empty lines between the fields are not removed, but still show up at the end of the files. But that could be corrected fairly easily if needed with sed, or perhaps grep -v.

Edit: You can remove the trailing blank lines with this after you're done:

Code:

for fname in xx* ; do
	printf '%s\n' '/^$/d' 'w' | ed -s "$fname"
done

Very slick, and both are very viable options.

Thanks again to you both for your help.