I wrote:
> Martin had originally proposed (2), which I rejected on the grounds
> that we don't treat multi-dimensional arrays as arrays-of-arrays for
> any other purpose.
Actually, after poking at it for awhile, that's an overstatement.
It's true that the type system doesn't think N-D arrays are
arrays-of-arrays, but there are individual functions/operators that do.
For instance, @> is in the its-a-flat-list-of-elements camp:
regression=# select array[[1,2],[3,4]] @> array[1,3];
?column?
----------
t
(1 row)
but || wants to preserve dimensionality:
regression=# select array[[1,2],[3,4]] || array[1];
ERROR: cannot concatenate incompatible arrays
DETAIL: Arrays with differing dimensions are not compatible for concatenation.
Various other functions dodge the issue by refusing to work on arrays
of more than one dimension.
There seem to be more functions that think arrays are flat than
otherwise, but it's not as black-and-white as I was thinking.
regards, tom lane