Shell Programming and Scripting

The variable "NF" holds the number of fields on the current record. Is there a variable that returns the current field within the record being processed? Given a record of "n" fields (variable number of fields for a given record), how can you tell which field you are actually on when you match a pattern?

I think NR does.

$ man awk

...

NR ordinal number of the current record

...

NR will give me the current record number. When Awk grabs a record, NR will surely give me the record number, but NF is set to the total number of fields in that record. I was hoping there was a variable that was also set to give me the current field that was being read. But as I think about it more, I guess this isnt really needed since I can iterate through all fields using a "for loop" and the value of NF. Thanks.

Hi,

I'm sure somebody will correct me if I am wrong here but I don't think awk loops through the fields on each record and therefore there isn't a current field as such.
The main input loop is executed for each record and references to fields are made as $1 = first field, $2 = second ... $NF = last field.
If you are matching a pattern without specifying the field in the match then you are basically asking if the pattern exists in the entire record ($0).

As you said - if you want to find exactly which field matches you would have to loop through the fields.

Matt.

AWK, as you know, is a line editor.

So it reads in a complete line and assigns variables $1 $2 $3 ... $NF which you can use in your awk statement for formatting your output or for calculations based on the variables you pick.

Based on this definition, your question is not valid. Awk processes all of the line at once not each individual variable one after the other.


NF is number of fields in a given record.

It can be used to print the field that happens to be the last one in the record.


NR counts the number of records

NR=2 is record #2.


IMHO, A great AWK book is the original one, the little grey book "The Awk Programming Language" by Aho/Kernigan/Weinberger...

If you want to find out which field you have matched then you would need to loop, e.g....

awk '/regexp/{for(x=1;x<=NF;x++)if($x~"regexp")print $x}' file1

Also consider the match function which gives the position of the matched pattern e.g...

awk 'match($0,"regexp"){print substr($0,RSTART,RLENGTH)}' file1