Does anyone think this is the correct behavior?
adaps_db=# select * from upc_usage order by 1 ; usage
----------53EABCCYPHER_GENERALH66HAWK_JOERSRAS61S65S70S76S92XWING
(14 rows)
It appears to be ignoring the underscore!
Database has LATIN1 encoding and was recently migrated from 8.0 to 8.3.