Shell Programming and Scripting

In this following command:
awk 'BEGIN{ORS=""}1'

what does '1' signifies that comes after closing curly brace '}' of awk? I guess, it does not mean 'first occurrence' because I verified that.

And, pls tell me how to override or suppress awk's field variables like $1, $2.. by positional parameters of the shell inside the script?

Thanks
Ibrahim

as far as i know..
if you include 1 it prints whatever in its pattern space..
pattern space is the place where it keeps the line or record its read..

not sure but... not able to recall..

An AWK program consists of a sequence of pattern-action statements (and optional function definitions)

Code:

pattern   { action statements }

The pattern is evaluated and if the result is true the action statements are executed.
If no pattern is not specified, the default value is true so all records are selected and the actions are allways executed.
For example the following code prints all input records converted to lowercase :

Code:

 { print tolower($0) }

If the action statements are not specified the dafault is print the input record.
For example, the following code prints all input records containing the string "OK" :

Code:

$0 ~ /OK/

In your code there are two sequences of pattern-action

Code:

BEGIN {      # Special pattern, allways true before anay of the input is read
   ORS=""    #  Set the Output Record Separator to null string
}            #
1            # Pattern value 1 is evaluated to true, so all records are selected
             #   The defalt action is 'print $0'
' inputfile

Your awk command merge all records (all records are printed without record separator because ORS="")

Jean-Pierre.

I still could not understand about the 'pattern value' or 'pattern space' concept. What happens if we change 1 as 2 in the above awk command?

it works only for 1 not for 2 or more..

The pattern is an expression that is evaluated for each record. If the result is true the action statements are executed.
If the result of the evaluation is a string or a number, the empty string "" and the number zero are considered to be false, all other values are true.
So the patterns 1 and 2 gives the same result that is true.

Jean-Pierre.

If I want to remove the second column as well as the delimiter (which is comma ',' here) following it, I am using this script.

Code:

echo '1,2,3' | awk 'BEGIN{FS=OFS=","} {$2=""}1'| awk '{sub(",,",",")}1'

output: 1,3

Is it possible to have a single awk command in the above instead of spawing 2 awks to accomplish this task?

Also, pls someone clarify me, which is the proper syntax of awk's sub()? sub(/<pattern>/,"<replacement>") (or) sub(<pattern>,"<replacement>")