John Gray wrote:
>
> Firstly, I appreciate this may be a hare-brained scheme, but I've been
> thinking about indexes in which the tuple pointer is not unique.
>
> The reason for my interest is storing XML documents in text fields in the
> database. (It could also help with particular kinds of full-text search?)
AFAIK this is what is known as an inverted index. This type of index is
most
often used in full-text indexes.
Something of similar nature is realised for "sets of integers" using
GiST
indexes and is available as "intarray" in contrib.
-------------------
Hannu