Storing some data to a different DB

I’m working on an Business Intelligence application where customers will
be adding values about different indicators. For example “Blue cars sold
daily”, then every day a record is stored representing that indicator.

So there can be many indicators and many many more values for each
indicator.

Right now everything is living in the same DB. Since the amount of
values can be huge after some time I’m now wondering if the values
should be stored in a different DB somehow.

I imagine the auto-increment IDs might someday restart and cause a nasty
issue if IDs that are used for the associations of the other objects get
overwritten.

Should I be worried about this and move the values storage to a separate
space? Any good suggestions?

Thanks.

On 26 November 2012 08:03, comopasta Gr [email protected] wrote:

I’m working on an Business Intelligence application where customers will
be adding values about different indicators. For example “Blue cars sold
daily”, then every day a record is stored representing that indicator.

So there can be many indicators and many many more values for each
indicator.

Right now everything is living in the same DB. Since the amount of
values can be huge after some time I’m now wondering if the values
should be stored in a different DB somehow.

What do you mean by huge?

I imagine the auto-increment IDs might someday restart and cause a nasty
issue if IDs that are used for the associations of the other objects get
overwritten.

That depends on what you mean by huge.

Should I be worried about this and move the values storage to a separate
space? Any good suggestions?

Again it depends on what you mean by huge.

Colin

Colin L. wrote in post #1086428:

That depends on what you mean by huge.

Huge to me is the number that will give me problems, and that is the
biggest number to which a DB can auto-increment before restarting the
IDs (or just failing)

Maybe I should start from the other end and find out that number for
typical DBs. Right now I’m on a Heroku Postgresql. I found some nice
discussion about this limit for SQL:

But I mixed one thing. Of course the auto-increment ID is independent
for each table. An the tables that will grow “hugely” (hopefully) are
not the center of an association, I select records from those 2 tables
using some fields and none of them is really the ID that the DB asigns
automatically.

I think I can relax for a while.

Thanks for your reply Colin.