IE> As for Sean's request about doing an efficient intersection, as
Alex
IE> was explaining there are only two ways to do this efficiently:
no, these are two ways to do it inefficiently, there is only one way
to do
it efficiently -- via combined index :)
True, but his request was efficienc
IE> The string tuple-sorting hack makes me cringe.
yep, this looks weird. but i mostly have id-and-time situation, and it works
fine for it.
IE> I know that I could make BDB and the lisp side sort lists based on
IE> their constituents pretty easily if you felt we could make the same
IE> ha
On 4/4/08, Ian Eslick <[EMAIL PROTECTED]> wrote:
> However, premature optimization often causes more trouble than it's worth.
> The easiest thing to do to get going is to index both slots and say:
>
> (intersection (get-instances-by-range 'event 'date )
> (get-instances-by-value 'e
Hi there,
The string tuple-sorting hack makes me cringe. I know that I could
make BDB and the lisp side sort lists based on their constituents
pretty easily if you felt we could make the same hack work in
postmodern.
(cons "Fred" 23) < (cons "Fred" 25) < (cons "Sally" 10)
Maybe we could
SR> What is the best/easiest/most-elephantish way to retrieve all events
SR> in the btree which have the 'lisp' tag and whose date falls on today?
so you have lots of possible tags and many tags per event?
then you need a btree with entry for each event for each tags, that is
( -> event_id) btr
??>> i believe that SQL RDBMS work this way too -- if one needs fast
??>> retrieval by several keys, he should create index on them. RDMBS knows
??>> how to sort tuples, though
SR> Well yes, but generally SQL RDBMS's will make efficient use of indexes
SR> it they are created on all the keys
> I'm really just curious about using
> 2 btree indices together without requiring the creation of a 3rd.
Of course I could be working against elephant while trying to do all of this.
So here is what I am trying to achieve, any pointers would be great.
Given the following code.
>>
(defpackage
On 4/4/08, Alex Mizrahi <[EMAIL PROTECTED]> wrote:
> (defclass test-event ()
> ((date :accessor date-of :initform (now) :index t) ...
> and it will create indices for you automatically.
Yes, although this is just an example of what I'm trying to achieve
and isn't exactly what i'm doing. I'm re
SR> (defclass test-event ()
SR> ((date :accessor date-of :initform (now))
SR>(name :accessor name-of :initform nil))
SR> (:metaclass persistent-metaclass))
SR> (defun create-date-indexer (a b c) (values t (date-of c)))
SR> (defun create-name-indexer (a b c) (values t (name-of c)))
SR> I'm hoping that someone can point me in the right direction here.
SR> Given 2 secondary indices on an indexed btree is it at all possible to
SR> create/use a derived index which is the union of the 2 created indices
SR> without requiring the explicit creation of a 3rd index
i'm not sure
10 matches
Mail list logo
