Shell Programming and Scripting

Hello,
I have an input file that looks like so:

Code:

 1 2 3
4 5 6
7 8 9

and I just want to print the first and third column (note: my actual file contains many many more fields so I don't want to use '{ print $NF }' for each field I want.

I tried using:

Code:

awk 'BEGIN {FS=" "} { for (i=1; i<=NF; i++) if (i<2 || i>2) print $i }' input.txt

but I do not get the output I want.

Any help is appreciated, Thank you.

Hello.

Me no understand.

If you want the first and third columns, why not use $1 and $3, instead of a for-loop?

Code:

$ echo "1 2 3
> 4 5 6
> 7 8 9" | awk '{print $1, $3}'
1 3
4 6
7 9
$

There's a bunch of other stuff to consider. Such as, the "input" you show starts with a leading space (presumably it's a space), and you set the field separator to a space, which means "1" would be the second field. Also, if you want the 1st and 3rd fields, $NF has no role to play. Also, you say the input file "looks like this", then proceed to claim that actually, it doesn't, etc.

Hi,
What I meant was that my actual file has the same format, but is much larger. It contains many more fields. I just put in three for the sake of simplicity.

There should be no space at the beginning, it should look like:

Code:

1 2 3
4 5 6
7 8 9

Again, my actual file contains many fields that are space separated, not just three. Thats why I wanted to use the for loop.

Hi.

That's fine.

But it's the same answer: if you want to extract specific fields, just use the field numbers. There's no need for a loop.

Code:

awk '{ print $1, $3 }' file

The tab and the space are the default field separators, so there is no need to set FS if your file is space-separated (doing so (FS=" ") would mean every space adds a new field).



edit: Actually, that ("(doing so (FS=" ") would mean every space adds a new field)") is not true, at least in GNU Awk

Code:

(two spaces between A, B and C...)
$ echo "A  B  C" | awk -F" " '{print $1, $2}'
A B
$ echo "A  B  C" | cut -d" " -f1,2
A

Try this as a starting point:

Code:

awk 'NR==1  {MX=split(COLS,C)} {for (i=1; i<=MX; i++) printf "%s ", $(C[i]); printf "\n"}' COLS="1 3" file
1 3 
4 6 
7 9

As with many of Rabu's threads, the problem specification isn't really clear.

Scott is absolutely correct that a for loop is not needed with the specified problem (and would be inefficient if the goal really is to always print the 1st and 3rd fields from every input line).

But, if the real goal is to print every field except the 2nd field, then a for loop makes sense. But using print instead of printf (without a \n in the format string) would create separate lines for each printed field instead the desired single line of output.

Scott,
The latest update to the POSIX standards says this about a single character ERE being used for FS:
Note that this is why FS="." and FS="*"work without needing to use FS="[.]" or FS="\*" when the goal is use a period or asterisk as a field separator without having to worry about them being interpreted as a metacharacters. But, if you want to use period and asterisk as field separators you need something like FS="[.]|[*]" instead of just FS=".|*".

Rabu,
Since {print $1,$3} does exactly what you have requested, please clearly explain why you don't want to use it (or clarify what you are trying to do so we understand why this code is not the best solution for your problem).

Don Cragun, you sir, are awesome.