Using variable inside awk

Shell Programming and Scripting
1

Hi,

Please help me how to use variables inside awk in code below:

ll | awk -v "yr=`date '+%Y'`" -v "mnth=`date '+%m'`" -v Jan=1 -v Feb=2 -v Mar=3 -v Apr=4 -v May=5 -v Jun=6 -v Jul=7 -v Aug=8 ' !/^d/ { if(NR>1) {printf "%-29s\t\t%s\t%5s\t\t%s %s,", $9,$1,$5,$`$6`,$7} }'

Thanks.

2

You already appear to be doing so, but not doing anything with the variables you set. What are you actually trying to do?

3

Manubatham, per the awk man page, the variables that you supply with the -v arg, are available in the BEGIN section only:

Here is how you can use it:

awk -v "Jan=1" 'BEGIN { print Jan }'

This will produce the output - 1, which is the value of the variable Jan.

4

Hi Corona,

Here you can see that $6 will return the month... that will be in format of Jan, Feb, Mar, or ... Now I want to convert month to numeric. Like 1 for Jan, 2 for Feb, 3 for Mar...so on.

Regards,
Manu

5

passing -v Jan=1 makes the variable named Jan contain "1". You don't get the string "Jan" anywhere.

How about:

awk 'BEGIN { split("Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec", MON, "|"); }
{ (rest of your code goes here) }'

That gives you an array MON where MON[1] gets you "Jan", MON[2] gets you "Feb", MON[$6] gets you the name of whatever number $6 is, and so forth.

[edit] Got it backwards. Hmm, thinking on it.

6

Am I using $Aug in correct way?

ll | awk -v "yr=`date '+%Y'`" -v "mnth=`date '+%m'`" -v Jan=1 -v Feb=2 -v Mar=3 -v Apr=4 -v May=5 -v Jun=6 -v Jul=7 -v Aug=8 ' !/^d/ { if(NR>1) {printf "%-29s\t\t%s\t%5s\t\t%s %s\n", $9,$1,$5,$Aug,$7} }'

Regards,
Manu

7 
awk 'BEGIN { split("Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec", MON, "|"); for(N=1; N<=12; N++) I[MON[N]]=N; }
{ (rest of your code goes here) }'

Now you have an array MON where MON[1]="Jan", and an array I where I["Jan"]=1.

---------- Post updated at 12:11 PM ---------- Previous update was at 12:10 PM ----------

You've already been told "no". $Aug will pick a different field depending on what number Aug is. Variables don't work that way in awk. print Aug; will print 8. Print $Aug will print field number 8, whatever it is. I don't think you can solve your problem with -v since you can't go from a number to a variable name.

You can use a number as an array index, which is what I have done. Try my above solution, which builds arrays out of a string of months.

8

Trying your solution, but $Aug always comes 8, I will never ask that question, see the output below:

ll | awk -v "yr=`date '+%Y'`" -v "mnth=`date '+%m'`" -v Jan=1 -v Feb=2 -v Mar=3 -v Apr=4 -v May=5 -v Jun=6 -v Jul=7 -v Aug=8 ' !/^d/ { if(NR>1) {printf "%-29s\t\t%s\t%5s\t\t%s %s\n", $9,$1,$5,$Aug,$7} }'

a.out                                   -rw-------          0           06:55 5
dir_str.txt                             -rw-------         35           10:23 19
dont_make_test                          ----------          0           10:03 23
get_dir_stru.sh                         -rwx------        599           11:02 5
prod_test                               -rwx------        521           13:27 23

Thanks,
Manu

9

$Aug will not come up 8. Aug will come up 8.

The -v solution will never work because variables don't work that way. You can't feed awk 8 and have awk decide "oh, there's a variable named Aug holding 8, you must mean it". It doesn't search for other variables with the same contents.

I may have stated it backwards, I["Jan"] will be 1 and MON[1] will be "Jan". Otherwise my solution does what you want.

1 Like
10

Hi Corona,

My current problem is related to the post

http://www.unix.com/shell-programming-scripting/165871-listing-files-particular-format-2.html

From there you can understand why I need this... and I am not getting how to implement your code in my solution. I have used $6 for month in mon format, but I want to convert it into numeric.

Thanks.

---------- Post updated at 01:15 AM ---------- Previous update was at 12:57 AM ----------

Thanks man.

Just clicked what you want to tell.

I have made my code like below--

!#/bin/sh
ll | awk -v "yr=`date '+%Y'`" -v "mnth=`date '+%m'`" 'BEGIN { split("Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec", MON, "|"); for(N=1; N<=12; N++) I[MON[N]]=N; } !/^d/ { if(NR>1) {printf "%-29s\t\t%s\t%5s\t\t%s %s,", $9,$1,$5,$6,$7}; if(index($8,":")==3) { if(I[$6]>mnth) { print yr-1 } else { print yr }} else { print $8 } }'

Do you want to suggest any change in the code above?

Thanks a lot. :slight_smile:

11

The skills involved are 'copy' and 'paste'.

awk 'BEGIN { split("Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec", MON, "|"); for(N=1; N<=12; N++) I[MON[N]]=N; }
 !/^d/ { if(NR>1) {printf "%-29s\t\t%s\t%5s\t\t%s %s\n", $9,$1,$5,I[$6],$7} }'

I don't know what you're trying to do with the code you crossposted.