2131100 COVS/100C E|FWP AREAS|MIDWEST|
2131100 COVS/100C E|FWP AREAS|CONSTRUCTION|
2131100 COVS/100C E|FWP AREAS|FTC|
2131100 COVS/100C E|FWP AREAS|FWP AREAS|
2131100 COVS/100C E|FWP AREAS|CLEANING|
2131100 COVS/100C E|FWP AREAS|FCRE|
2131200 COVS/100C UPS|UPS ROOMS|CONSTRUCTION|
2131200 COVS/100C UPS|UPS ROOMS|UPS ROOMS|
2131200 COVS/100C UPS|UPS ROOMS|FCRE|
2131200 COVS/100C UPS|UPS ROOMS|EMERGENCY|

I am running a report that produces the output shown above. I am trying to use awk in the script to format the output to look like this

2131100 COVS/100C E|FWP AREAS|MIDWEST|CONSTRUCTION|FTC|FWP AREAS|CLEANING|FCRE|

All information would be on one line but the carriage return was forced when typing into the area for this post.
Basically, I want to keep only the first occurence of the information in colums 1 and 2. I then want to list the information in column 3 from left to right in the same row. I want to do this for each new number listed, ie: one row that looks like the example above for 2131100 another row for 2131200 etc...

Attached below is a copy of the AWK script which does not seem to be working properly.

Any insight on how to accomplish this would be greatly appreciated.

awk -F"|" 'BEGIN {SS_A=""}
{if (NR==1)
{SS_A=$1; printf $0}
else
{if (SS_A==$1)
{printf "|"$2}
else
{SS_A=$1;
printf "\n"$0}
}
} END {printf "\n"}' $SF_OUT

The following works for me:

#!/usr/bin/awk -f

BEGIN {
FS="|"
line=""
}

$1 $2 == line {
printf $3 FS
next
}

{
if (line != "")
print ""
printf $0
line=$1 $2
}

END {
print ""
}