Submitted by: Christian Sigg; Assigned to: Nobody; R-Forge link

It seems that duplicate column naming is broken in 1.9.3. It used to be that data.table appends counters to duplicate column names after joins. For example

A <- data.table(foo = 1:2, bar = 3:4)
setkey(A, foo)
A[A[A]]

resulted in columns foo, bar, bar.1 and bar.2. Now I get foo, bar, i.bar and i.bar (again), which prevents accessing the proper bar column by name in later queries. Is this intended behavior or a regression?

Thanks,
Christian