Evandro's mailing lists (Please, don't send personal messages to this
address) wrote:
I'm doing a PhD in data mining and I need more than 1600 columns. I got an
error message saying that I can not use more than 1600 columns.
It is happening because I have to change categorical values to binary
creating new columns.
Perhaps you don't want a relational database at all if you are
stretching it to match your client application in this way. Do I have it
right that you have something like
Table: bird_sighting_facts (bird, category, value)
1 | wingspan | 120mm
2 | beak-colour | red
3 | chest-colour| blue
...
And are converting it into:
expanded_bird_facts (bird, cat_wingspan, cat_beak_colour,
cat_chest_colour, ...)
In which case since you'll almost certainly be throwing away any
relational integrity you had in the first case I'd just throw a
lightweight wrapper around some dbfile files or similar.
--
Richard Huxton
Archonet Ltd
---------------------------(end of broadcast)---------------------------
TIP 4: Have you searched our list archives?
http://archives.postgresql.org
