Liveware Publishing Support Forum

I have got a weird issue with R&R 9 & 10. I^m running Oracle as the backend with Oracle ODBC. This is probably a issue for tech support and I^m working on getting some money for that. I^m just curious if anyone else has any ideas.____When developing a new report, everything is fine. (by fine, I mean, response time is reasonable.) But try and open up a previously developed report, and, well, extremely slow is an understatement. Once the report gets back from outer space, the report acts normally. The RDBMS is cpu bound during these periods. Reports with no joins are fine.____I^ve done the usual swap the layers and versions. I^ve patched everything I can safely do. I am at the point of capturing the queries as they execute in the database. ____There seems to be a sequence of queries that R&R uses for initialization. These queries are of the form "select * from table1 where 0 = 1" where table1 is the name of each table in the report. These are called sequentially, one for each table. (table1, then table2, then table3, etc.)____Then, the whopper query hits, "Select * from table1, table2, table3 __where 1 = 2" (it^s hard not to shudder at this point). It looks as though the joins from the where clause are missing! Now, these tables are 100K rows each so like wow, no wonder.____I^m thinking that the query where clause should include the "joins" with a "and 1 = 2" tacked on.____Has anybody else run into this?

With kathleen^s pointer that I should look at the RRWSQL.INI file, I have managed to narrow this startup issue down to two functions in the generic odbc section at the bottom of the file. If you comment out date = curdate() and time = curtime() lines, startup performance is restored. ____I have plenty more testing to do but it looks good so far.____-Rob

That is certainly an interesting finding.______Kathleen__R&R Support

RRWSQL.INI should be RSWSQL.INI____Sorry about that...