Greeting, Lutz!
Please don't top-post on the PG mailing lists, our style is to relpy
in-line.
* Lutz Fischer (l.fisc...@ed.ac.uk) wrote:
> I often need some data from [s] where I don't care about [sp]. So in
> how far does having these arrays a part of [s] would make these
> queries slower. Or wo
Hi Stephen,
Thanks for your reply. The data in the sub table (sp) are only read in
as a block. Meaning I will always read in all entries in [sp] that
belong to one entry in [s]. Meaning I would not lose much in terms of
what I could do with the data in [sp] and I could be saving around 2.8K
Greetings,
* Lutz Fischer (l.fisc...@ed.ac.uk) wrote:
> Data in [sp] are never changed. I can probably reduce the size by
> changing datatypes from numeric to float but I was wondering if it
> would be more efficient - primarily in terms of storage - to change
> the structure to have two arrays i
Hi,
I have two tables
s {
id bigint NOT NULL PRIMARY KEY,
...
}
sp {
id bigint PRIMARY KEY,
sid bigint REFERENCES s (id),
i numeric,
m numeric
...
}
I have for each entry in [s] on average around 120 entries in [sp]. And
that table has become the largest table in my
