[SOLVED] How to run code on a certain line number in AWK

MTK358 · 05-09-2010, 08:25 AM

I want to be able to run a piece of AWK code on a record specified by number, not by a regex.

This is my current solution, but i think there must be a better way:

Code:

{
        if (NR == 3)
        {
                # do stuff
        }
}

Is it possible, and how is it done?

ntubski · 05-09-2010, 09:03 AM

Code:

 NR == 3 { # do stuff }

colucix · 05-09-2010, 09:14 AM

As ntubski said! Take in mind that rules in awk have the following format

Code:

pattern { action }

where pattern can be a regular expression or a comparison expression. If the pattern is not specified, the action is performed on every line (record) of the input file. If the action is not specified, the default one (print $0) is applied.

MTK358 · 05-09-2010, 09:37 AM

And will this work for a range of records?:

Code:

(NR >= 3 && NR <= 8) { print }

EDIT: I tried it and it does work.

colucix · 05-09-2010, 09:40 AM

Yes! And parentheses are not necessary:

Code:

NR >= 3 && NR <= 8 { action }

MTK358 · 05-09-2010, 09:41 AM

I didn't know that you can put something other than a regex to match.

Is it possible to match something not even related to the current record?

colucix · 05-09-2010, 09:50 AM

Quote:

Originally Posted by MTK358

Is it possible to match something not even related to the current record?

Well, not really. Built-in variables as NR, NF, and so on are related to the current record. Regular expression are matched against the current record. However you can test for variables that have been assigned before, for example:

Code:

/regexp/ { count++ }
count >= 10 { action }

In this poor example, the variable count is incremented every time a record matches the regexp, whereas the second rule is executed only after count has reached a value of 10. In this case the answer is yes: the value of count is not strictly related to the current record.