Re: Uniquely identifying a record in SQL without a unique key?

I've been largely silent on this thread since I started it, because it took me all weekend to formulate an intelligent next question to ask.

So. Am I to understand that a "result set," while it may look at times like simply a collection of records, is in fact a combination of a cursor, and the meta-data of the file (or at least, of the fields in the select), and perhaps some other things?

The primary objective here is to expand an existing database access utility to cover databases other than the local DB2, with a secondary objective of perhaps giving the user the option of going through JDBC (via JDBCR4) for files containing data types we only have limited access to through native RLA (e.g., datalinks), or no access at all (e.g., CLOBs and BLOBs).

So we want to be able to display records without acquiring update locks on them, but to be able to go back into any one those records with absolute certainty that it's the right record, acquire an update lock, and update it.

--
JHHL