Knowing my volatile memory, I think it is best that I take you up on your suggestion of adding the explanation and examples in the javadoc. I hope to have a patch with these comments by early tomorrow.
thanks, Mamta On 5/10/05, Daniel John Debrunner <[EMAIL PROTECTED]> wrote: > Mamta Satoor wrote: > > > Hi Dan, > > > > Here is another patch with more comments on the getTable/SchemaName > > and getSourceTable/SchemaName methods. > > > > As for your following question > > "Then with this patch ColumnReference has a getSchemaName() and a > > getSourceSchemaName(). I'm unclear on what the intended behaviour of > > each is, and why we need two methods. If getTableName() means return the > > "user" name or correlation name, then logically getSchemaName() would > > follow the same pattern, but user names or correlation names don't have > > schemas." > > You are right that there is no way to provide a correlation name for > > schema. But a user sql can look like following > > select app1.t1.c1, app2.t1.c1 from app1.t1, app2.t1 > > For this sql, getSchemaName(and getSourceSchemaName) will return app1 > > for first column and app2 for second column. > > > > For select t1.c1 from t1, getSchemaName for first column will return > > null but getSourceSchemaName will return app (assuming we are in the > > schema app). > > > > Aggregate functions rely on ColumnReference.getSchemaName at generate time. > > The above comments and examples are really helpful, and would be even > more helpful if they were added to the javadoc comments for these > methods. :-) > > A third example would be 'select a.c1 from t1 a' > > I'll work on committing this patch. > > Dan. > >
