From: Josh Cheek <josh.cheek@...>
Date: 2009-09-15T03:00:38+09:00
Subject: Re: Pre-allocate large amount of memory?

--000e0cd5a1c0a94ce404738d5cfe
Content-Type: text/plain; charset=ISO-8859-1

On Mon, Sep 14, 2009 at 12:37 PM, Carsten Gehling <carsten@sarum.dk> wrote:

> Robert Klemme wrote:
> >
> > Maybe it's not such a good idea to read everything in memory. After
> > all, retrieving large volumes of data from disk is where RDBMS's
> > shine.
>
> My reason for doing this is simply to avoid multiple sql-queries in a
> recursive self-referencing datastructure. Usually this can be avoided by
> querying the entire subset of data needed, and manipulate it in memory.
> Unfortunately the data in question does not enable me to just query the
> subset of data.
>
> I will definetely have a look at Redis. Seems promising.
>
> Thanks.
>
> - Carsten
> --
> Posted via http://www.ruby-forum.com/.
>
>
I had a situation like that, where I ended up using one of the nested set
implementations (I think it was Awesomer Nested Set) instead of a tree
hierarchy. I ended up having to write my own (very ugly) iterating
interpreter which analyzed each node in to figure out it's recursive
context. It took something like 3 or 4 blocks for the different places you
could inject code based on a recursive interpretation, but it did get me
down to a single database call, and it did only require a single pass
through the data. At one point (debugging, ahem) I was very strongly
considering going with the approach you are taking. Good luck with it, and
let me know how it turns out, I might opt for that approach instead, next
time.

--000e0cd5a1c0a94ce404738d5cfe--