[Haskell-cafe] efficient data structure: column with repeating values

Is there some good data type out there that basically provides a simple table, but with optimization for repeating values on one column? Something like: Data Table a b ...where it assumes that 'a' values will usually be unique, while 'b' values will usually be repeated from a small set? (But not needing to be fixed beforehand.) Like in... client | patron --------------- Bob | Tom Sarah | Tom Dick | Tom George | Harry Moe | Harry -- frigidcode.com