Lee Hachadoorian <lee.hachadoorian@xxxxxxxxx> writes: > How is the number of columns in a join determined? When I combine somewhere > around 90 tables in a JOIN, the query returns: > ERROR: joins can have at most 32767 columns It's the sum of the number of columns in the base tables. > I'm sure most people will say "Why the hell are you joining 90 tables." Not only that, but why are you working with over-300-column tables? Seems like your schema design needs rethinking. > I've asked this list before for advice on how to work with the > approximately 23,000 column American Community Survey dataset, Are there really 23000 populated values in each row? I hesitate to suggest an EAV approach, but it kinda seems like you need to go in that direction. You're never going to get decent performance out of a schema that requires 100-way joins, even if you avoid bumping up against hard limits. regards, tom lane -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
