Question : How can I grep, parse lines, and return count by certain fields?

-need a one liner that greps a file for all entries with 'FAIL'
-then it will parse through the lines and return a count for each unique name in field 'd' (note that fields are not always in the same order)
-if field d does not exist, count by field 'g'
Example:
XXXXXXXXXXXXXXXXXXXXXX$a=1234:d=fred:y=alphazXXXXXXXXXXXmFAIL
XXXXXXXXXXXXXXXXXXXXXX$d=fred:a=4321:y=bravozXXXXXXXXXXXmFAIL
XXXXXXXXXXXXXXXXXXXXXX$a=1234:g=jackson:y=charliezXXXXXXXXXXXmFAIL
XXXXXXXXXXXXXXXXXXXXXX$y=delta:a=1357:d=bobbyzXXXXXXXXXXXmFAIL
--------------------------------
If you were to run the one-liner on the above lines, the output should be:
fred 2
jackson 1
bobby 1

-If need be, it is ok to call perl for this execution

Answer : How can I grep, parse lines, and return count by certain fields?

zgrep FAIL file | perl -ne '(/\bd=(\w+)/||/\bg=(\w+)/) && $c{$1}++;END{print"$_ $c{$_}\n" for keys %c}'