I am only hoping that some point of time the bug reported by me will get reproduced from your end and you will find time to work on this.
I can assure you that absent further feedback and details from you no one will be working on this. Hundreds of thousands of people have hundreds, if not thousands, of tables, in production databases, running 24x7x365, and do not experience this behavior. The problem is specific to your organization and thus your organization's responsibility.
David J.
Hi,
What you are telling to us is not a simple bug. It's something that would make Postgres completely unusable as a database. Luckily nobody experienced it so far except you.(Mind you, people on this list have countless hours of experience building, running and maintaining Postgres databases and I don't even dare to guess how much data that means.)
It is analogous with telling the Java core developers that if statement doesn't work correctly, it messes up the evaluation of the condition sometimes.
It's highly unlikely. And this is the only thing we can tell you without further information about your issue.
So if you want our help, please provide us at least the table definition and the select statement producing the wrong result.
On top of that log and check the dml statements on your database to see if something modifies the data without your knowledge.