On 2010-12-08, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Adrian Klaver <adrian.klaver@gmail.com> writes:
>> On 12/08/2010 08:04 AM, Tom Lane wrote:
>>> The rationale for having a limit of this sort is (a) we *don't* want
>>> the upper limit of declarable length to be encoding-dependent; and
>>> (b) if you are trying to declare an upper limit that's got more than a
>>> few digits in it, you almost certainly ought to not be declaring a limit
>>> at all.
>
>> Well that explains it :) Would it be possible to change the below
>> section in the docs to state that the declared max value of n is limited
>> to a max string size of 10Mb?
>
> I don't really see any point in that. The value is meant to be an order
> of magnitude or so more than anything that's sane according to point (b).
> If you think you need to know what it is, you're already doing it wrong.
I have some values of perhaps 20Mb that I might want to store samples
of in a partitioned table. (so I can delete them easily) what's the right way?
I guess I could just keep them as disk files and rotate the
directories as I rotate partitions.
--
⚂⚃ 100% natural