Here's the patch. It's fairly trivial IMO, but as the unittest (and sample in the above email) shows, it reduces LoC quite a lot if you want to do this sort of join and you know the table/column names.
2013/7/18 Henry Saputra <[email protected]> > Hi Kasper, > > You could just attach proposed patch here and review through list for a > while. I have pinged INFRA about the JIRA component. > > I will also try to get reviewboard (https://reviews.apache.org) for > MetaModel > > - Henry > > > On Thu, Jul 18, 2013 at 1:20 AM, Kasper Sørensen < > [email protected]> wrote: > > > Hi guys, > > > > I have a few improvements to the fluent Query builder API, that I would > > like to commit. > > > > Basically the fluent Query builder API supports building joins, but only > if > > you've already traversed Table and Column objects from the schema model. > > What I would like to improve is to add String-based builder methods so > that > > instead of something quite verbose like this... > > > > Table table1 = dataContext.getTableByName("table1"); > > Column col1 = table1.getColumnByName("col1"); > > > > Table table2 = dataContext.getTableByName("table2"); > > Column col2 = table2.getColumnByName("col2"); > > > > DataSet ds = dataContext.query() > > .from(table1).innerJoin(table2).on(col1, col2).execute(); > > > > > > One could alternatively just use the table/column names directly in the > > builder API, like this: > > > > DataSet ds = dataContext.query() > > .from("table1").innerJoin("table2").on("col1", "col2").execute(); > > > > > > Since we dont have JIRA up and running yet, I didn't find a proper way to > > add this as an improvement anywhere. Should I just commit anyway, or hold > > my commit until JIRA is up? Or are there any objections to the > improvement? > > >
