Use dbplyr to handle more of our *_joins

NEWS.md

@@ -8,6 +8,10 @@
 
 * `update_snapshot()` has been optimized and now runs faster on all the supported backends (#137).
 
+* `*_joins()` have been more robust:
+  * `dbplyr` is now used internally which improves `full_join` and adds `anti_join` and `semi_join` (#157).
+  * When not supplying a `na_by` argument no input validation is made and unmodified `dplyr::*_join()` is called (#156).
+
 ## Documentation
 
 * A vignette including benchmarks of `update_snapshot()` across various backends is added (#138).