increment variable in awk based on two columns - linux

Question

I am writing an awk script that parses a CSV file, compares one column containing date, and another column containing activity type, and then prints the count of a particular activity.
The code I have written is:
NOW=$(date --date="5 days ago" +"%Y%m%d")
awk -F "," -v mydate=$NOW '{
var_1=1;
var_2=1;} {
if ( substr($8,2,8) == mydate ) {
if ( $6 == 1001 ) {
var_1++;
}
else if ( $6 == 1003 ) {
var_2++;
}
}
print var_1 var_2
}' *.csv
The output I get is
11
11
11
11
11
11
I believe the issue is something to do with the way I have defined var_1 and var_2; they are reinitialized or something.
Also I want to only print the final value of var_1 and var_2; at the moment, it gets printed with every iteration of awk.
Any advice?

Answer 1

You have two blocks that are executed on each line of data:
{ var_1=1; var_2=1; } which sets the variables to 1 on each pass.
{
if ( substr($8,2,8) == mydate ) {
if ( $6 == 1001 ) {
var_1++;
}
else if ( $6 == 1003 ) {
var_2++;
}
}
print var_1 var_2
} which prints the values of var_1 and var_2 as concatenated strings (hence no space between the 1 and 1).
It appears that either the substr() condition or the $6 condition is not being matched, ever.
You probably wanted BEGIN before the first block, but why you'd start at 1 rather than 0 is not obvious. If you started the counts at 0, you wouldn't need a BEGIN block. You should probably use print var_1, var_2 to separate the two values.
As for why the matches aren't matching, there's no way to say without any sample data to work on, but you could debug by printing out $8 and $6 for each line (and mydate, too; and maybe substr($8,2,8)), so you can see what is happening.
If you only want the values to print at the end, then (once you've debugged what's happening during the main action), you can place the print in an END block:
END { print var_1, var_2 }