Shell Programming and Scripting

I have a file with a structure like this:

Code:

Database sequence: some data
Database position: number
Query: identifier
Location: number
E-value: number

       0         .     :     .     :     .
           STRINGSTRINGSTRINGSTRING
           ||||||||||||||||||||||||||||
           STRINGSTRINGSTRINGSTRING

@

Database sequence: some data
Database position: number
Query: identifier
Location: number
E-value: number

       0     .     :     .     :     .
           STRINGSTRINGSTRINGSTRING
           ||||||||||||||||||||||||||||
           STRINGSTRINGSTRINGSTRING

@

etc.
repeated for many entries, each being unique.

I am trying to capture the first line of each entry, and the first of two matching strings (it actually doesn't matter if it's the first or second, but for precision I want the first).

I found that I can separate the records using the "@" and the fields using a newline, and then print the fields I want using sed (I would post a link but the rules won't let me, yet. Thanks, GNU awk User's Manual).

So I copied the example from there as example.awk:

Code:

BEGIN { RS = "@" ; FS = "\n" }

Code:

{

              print $1
              print $8
}

and I run: awk -f example.awk test.txt > output.txt.
In output.txt, I only get the first record. It looks like

Code:

Database sequence: some data
           STRINGSTRINGSTRINGSTRING

which is what I want, but of course I want them all.

Sorry for the dumb noob question. Thanks in advance.

You are printing fields 1 and 8, and getting fields 1 and 8.

To get all fields, just 'print'.

It looks like your record separator is an AT symbol on a line by it's self (perhaps also followed by one or more other blank lines?)

I'd suggest trying RS of @\n*:

Code:

BEGIN { RS = "@\n*" ; FS = "\n" }
{
    print $1
    print $8
}

Thanks. Corona688, I only want fields 1 and 8, but I want them for ALL records in the file. I'm only getting 1 & 8 for the FIRST record in the file.

Chubler_XL, you are correct there is whitespace/blank lines around "@" symbol but unfortunately with RS = "@\n*" , I still only get the first record.

I see what you mean. The extra newlines throw off what field your data is.

RS="@\n*" works fine for me. Please post input data it doesn't work for. [edit] Maybe there's carriage returns in there too? RS="@[\r\n]*"

Also, if you're on solaris, try nawk.

I don't know whether I mis-typed something earlier, but the RS = "@\n*" is working for me. Thanks so much!