For me, Option 3 first, then 1 if I can't use an SP for one reason or another. Actually, stuffing the strings-whether they are SP calls or raw SQL-into a CDATA node in an XML doc might be a nice option. Perhaps not terribly elegant, but you'd minimize messy character translation/special handling issues. Come up with a nice way to organize these (by classname/methodname maybe) in the XML doc and you'd have a pretty friendly and self-documenting way to pull these in at run-time.
Option 2 sounds like a real headache. -------------------------- How do you NOT sprinkle RDMS specific SQL dialects in your code? Option 1: Store all the strings in some resource and swap in the version you need at runtime Option 2: Build the SQL string from a "PhraseBook" by mapping your version of SQL to your DBMS version of SQL. Option 3: Use stored Procs for cause the differences are fewer? Others? Thanks again, Jim You can read messages from the DOTNET archive, unsubscribe from DOTNET, or subscribe to other DevelopMentor lists at http://discuss.develop.com. You can read messages from the DOTNET archive, unsubscribe from DOTNET, or subscribe to other DevelopMentor lists at http://discuss.develop.com.
