Yes, I'm sure. Early yesterday I did get duplicate key errors. That's when I looked on stackexchange to learn how to reset the sequence's max value to the value of the number of rows in the table. Not only did my attempt to add a single new company to the companies table change all company names in that one industry to the new name, but I just discovered that it changed all rows in that column to the new company name:
At present the belief there is a bug in PostgreSQL is unsubstantiated.
I suggest you look internally for how an update command that caused the resultant data could have been executed. That is much more plausible, and thus a better use of time, if you want to spend more time on this, than trying to produce the observed behavior again using just insert and setval(...) commands.
