Hacker Read

pgaddict · 2017-09-22 11:47:41+00:00

Yes, but partitioning also has a number of limitations that may be quite undesirable. Like inability to create UNIQUE constraints or primary keys, for example.

kjeetgill | karma 3922 | avg karma 3.2 · | 2019-05-21 19:29:04+00:00

We'd considered it, but partitioning hurts other things like turning a single row query into # of partitions look up.

simonask | karma 125 | avg karma 2.16 · | 2017-01-20 06:50:18+00:00

That's correct. I also believe you could in theory forgo partition tolerance, it just wouldn't be a very useful system. Maybe there are use cases out there for it.

abalone | karma 12497 | avg karma 4.68 · | 2017-06-12 22:19:39

Cool, yeah, that's what I meant by partitioning, which I guess is more of a database term.

erik_seaberg | karma 4377 | avg karma 1.56 · | 2019-11-30 15:20:45

Yeah, partition size is only limited if you create a local index, because those are strongly consistent.

jackcosgrove | karma 3623 | avg karma 3.38 · | 2020-04-28 15:06:32

Partitioning sounds like a great solution, thank you for this.

fernvenue | karma 194 | avg karma 3.34 · | 2023-08-07 09:01:25

Is there any reson that it has to be a partition?

noduerme | karma 11910 | avg karma 2.8 · | 2023-05-22 22:39:29

Any partitioning?

fuckadtech | karma 1 | avg karma 0.33 · | 2020-12-09 00:48:06

Yes, but this assumes that users have the foresight to over-provision (enough). It becomes an operational pain in reality.

There is also no getting around the variation of the noisy (or slow) neighbor problem slowing down an entire partition. That also becomes an operational pain.

reply

mcartyem | karma 186 | avg karma 1.16 · | 2012-06-05 20:19:12

Why would anyone partition tables -- ever?

ants_a | karma 244 | avg karma 1.72 · | 2020-08-12 20:04:32+00:00

Why wouldn't you use built in partitioning for this?

HPsquared | karma 7582 | avg karma 2.12 · | 2021-11-12 05:52:44

Is there any advantage to doing it this way? Partitioning?

aaronblohowiak | karma 7361 | avg karma 2.49 · | 2014-03-13 01:39:19+00:00

No, it is just a system that can lose data in the face of a partition.

politician | karma 6427 | avg karma 2.21 · | 2015-03-25 01:00:50+00:00

It exists, it's just that assuming partitions never happen isn't very practical.

biksnep | karma 2 | avg karma 2.0 · | 2017-05-11 01:42:38+00:00

Partition tolerance?

siliconviking | karma 336 | avg karma 2.69 · | 2013-08-17 13:32:43

Very interesting - I wasn't aware of this! Looks like this could provide a solid first cut to the partitioning process. Thanks for recommending...

Deejahll | karma 251 | avg karma 5.98 · | 2009-07-12 19:32:00+00:00

Ah, this is very helpful. I was aware that I wanted to do something like #2, but I was not aware that such a partitioning was the norm.

yread | karma 18099 | avg karma 4.2 · | 2011-06-23 13:20:36+00:00

Yes, check out the Partition I 12.1 table 6 of ECMA-335 standard

Gracana | karma 3569 | avg karma 2.83 · | 2017-07-11 16:42:28+00:00

That is an example of not considering your use case and failing, yes. :P

Flexible/resizable partitions are definitely easier, but I think you could have gotten the fixed scheme right if you had tried.

reply

rainmanwy | karma 1 | avg karma 0.2 · | 2020-01-11 03:15:04+00:00

I think could also try the partition feature.