Thread: Is there an upper limit on the size of an array if it is indexed?

I have created a unique index on an array of integers, and it works for a
small test case:

create table i(id int not null primary key,
v int[] not null unique );

insert into i(id, v) values(1, ARRAY[1,2]);

-- fails
insert into i(id, v) values(2, ARRAY[1,2]);
-- succeeds
insert into i(id, v) values(2, ARRAY[2,1]);

Is there an upper limit on the size of an array being indexed? I am
currently considering arrays of 2K-3K integers.




--
View this message in context:
http://postgresql.1045698.n5.nabble.com/Is-there-an-upper-limit-on-the-size-of-an-array-if-it-is-indexed-tp5793380.html
Sent from the PostgreSQL - sql mailing list archive at Nabble.com.

On 02/24/2014 01:30 PM, AlexK wrote:
> I have created a unique index on an array of integers, and it works for a
> small test case:
>
> create table i(id int not null primary key,
> v int[] not null unique );
>
> insert into i(id, v) values(1, ARRAY[1,2]);
>
> -- fails
> insert into i(id, v) values(2, ARRAY[1,2]);
> -- succeeds
> insert into i(id, v) values(2, ARRAY[2,1]);
>
> Is there an upper limit on the size of an array being indexed? I am
> currently considering arrays of 2K-3K integers.
>
>
>
What are the odds?  Might be worth the risk.

rjs

AlexK <alkuzo@gmail.com> writes:
> I have created a unique index on an array of integers, and it works for a
> small test case:

> create table i(id int not null primary key,
> v int[] not null unique );

> insert into i(id, v) values(1, ARRAY[1,2]);

> -- fails
> insert into i(id, v) values(2, ARRAY[1,2]);
> -- succeeds
> insert into i(id, v) values(2, ARRAY[2,1]);

> Is there an upper limit on the size of an array being indexed? I am
> currently considering arrays of 2K-3K integers.

PG btrees have an item size limit of a couple of KB, so I don't think
this is going to work well for you.

Some people have gotten around that by using a unique index on a hash
value.  If you use something like MD5, the odds of a false collision
are probably low enough to tolerate.
        regards, tom lane

Tom, is this limitation documented anywhere?



--
View this message in context:
http://postgresql.1045698.n5.nabble.com/Is-there-an-upper-limit-on-the-size-of-an-array-if-it-is-indexed-tp5793380p5793387.html
Sent from the PostgreSQL - sql mailing list archive at Nabble.com.