Awk: use variable defined in begin

UNIX for Beginners Q & A
1

Hi there,

I'm working with file more than 400K lines, 60 columns. Column count is going to be multiple of 12: 60, 12, 72 or so.
NF/12 gives me on how many iterations I've to do to check certain value.

For example: 7, 14th if only 24 columns in file.
7th, 14th and 21st if 36 columns in file.

I'd like to store and use NF/12 after calculating in BEGIN. But I can't use it.


head CHR_10_paste.info | awk ' BEGIN{ incr=NF/12 } NR>1 { print $incr }'

incr is 0 so this one-liner prints all the columns.
How do I use it later on once calculated in BEGIN?

2

Try:

print incr

awk does not use a $ sign to get the value stored in the variable..

3

It prints 0

4

Oww that is right,
Try:

FNR==1{ incr=NF/12 }

If NF is the same for every line in the file

In the BEGIN section NF is not yet initialized..

And remove the NR>1 condition

So:

awk ' FNR==1{ incr=NF/12 } { print incr }'
5

In the BEGIN section, no file has been opened and read yet, so NF can't be set. And, if I understand correctly, NF changes with every line read, doesn't it? So incr should not be calculated in the BEGIN section, but for every single line.

6

NF is going to be constant in a file for all the rows.
A file may have a different number of columns. But they are going to be consistent across all the lines.
I hope I'm making sense.

7

That's why Scrutinizer suggests to set the incr in line 1.

NR==1{ incr=NF/12 } { print incr, "line", NR }

A next skips the further processing of the current line.

NR==1{ incr=NF/12; next } { print incr, "line", NR }