Re: how to use partitions once created?
- From: "Rich" <Rich@xxxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Tue, 3 Jan 2006 19:52:02 -0800
Thank you very much again for your reply and for the explanations and
examples. Your example queries worked perfectly. I just have to keep doing
these queries until I see the pattern/logic of MDX queries and functions.
And yes, I partitioned by States - Conneticut and Vermont. Is that why this
query worked?
SELECT
{[Measures].[Amount],[Measures].[Sh Num]} ON COLUMNS,
{[Fund].[Fund Name].Members}
ON ROWS
FROM [Investments]
WHERE [Account].[State Abbr].[CT]
BTW, is a Where clause (in MDX) and a slice the same thing - semantically
speaking?
Thanks,
Rich
"Akshai Mirchandani [MS]" wrote:
> What you really want to do is:
>
> SELECT
> {[Measures].[Amount],[Measures].[Sh Num]} ON COLUMNS,
> {Crossjoin([Fund].[Fund Name].Members, { [Account].[State Abbr].[CT] } )}
> ON ROWS
> FROM [Investments]
>
> i.e. specify which member you want in the set itself. The WHERE clause in
> MDX is not meant for scenarios like this -- it is more for something like
> this:
>
> SELECT
> {[Measures].[Amount],[Measures].[Sh Num]} ON COLUMNS,
> {[Fund].[Fund Name].Members}
> ON ROWS
> FROM [Investments]
> WHERE [Account].[State Abbr].[CT]
>
> i.e you want all the funds but sliced for the CT state.
>
> Btw, you typically specify a slice as an MDX expression when you create the
> partition -- e.g. if you partitioned by states, it might be an expression:
> [Account].[State Abbr].[CT]
>
> Thanks,
> Akshai
> --
> This posting is provided "AS IS" with no warranties, and confers no rights
> Please do not send email directly to this alias. This alias is for newsgroup
> purposes only.
>
> "Rich" <Rich@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
> news:788FD303-6476-4EC3-9EED-30BE4871881A@xxxxxxxxxxxxxxxx
> > Thank you for your reply to my question. May I ask, how do you specify a
> > slice ona partition? For example, my cube is called "Investments". One
> > of
> > the partitions is called "Conneticut"
> >
> > SELECT
> > {[Measures].[Amount],[Measures].[Sh Num]} ON COLUMNS,
> > {Crossjoin([Fund].[Fund Name].Members, [Account].[State Abbr].Members )}
> > ON ROWS
> > FROM [Investments]
> > --Where
> >
> > This query yields something like (without any where clause):
> >
> > Amount Sh Num
> > 20th-Century Select CA $30,000.00 1,111.10
> > CT $38,000.00 1,421.62
> > FL
> > Acorn International CA
> > CT $245,841.30
> > FL
> >
> >
> > I tried adding something to the Where Clause
> >
> > Where [Account].[State Abbr].CT
> >
> > but got this error message "Unable to open cellset
> > Formula error - duplicate dimensions across (independent) axes - when
> > calculating a query axis"
> >
> > Well, even though this is probably not a partition slice, how could I get
> > this slice to work?
> >
> > Thanks,
> > Rich
> >
> >
> >
> > "Akshai Mirchandani [MS]" wrote:
> >
> >> Partitions are physical objects -- they allow better management of the
> >> data
> >> in a cube and in some ways improve performance. But you don't (and can't)
> >> query them directly -- the cube is the object that you care about.
> >>
> >> The "better management and performance" include:
> >> - creating rolling partitions so that old partitions don't need to be
> >> re-processed and new data can be incrementally added into the "current"
> >> partition
> >> - specifying slices on partitions so that partitions outside the slice in
> >> the query don't need to scanned
> >> - etc.
> >>
> >> Thanks,
> >> Akshai
> >> --
> >> This posting is provided "AS IS" with no warranties, and confers no
> >> rights
> >> Please do not send email directly to this alias. This alias is for
> >> newsgroup
> >> purposes only.
> >>
> >> "Rich" <Rich@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
> >> news:D8600CDE-4B62-4D37-BCA8-8C733EC5340A@xxxxxxxxxxxxxxxx
> >> > Hello,
> >> >
> >> > I have a data warehouse/datamart database (please correct me what the
> >> > correct name is for the database containing the cubes) containing 2
> >> > cubes
> >> > called "Dividends" and "Investments". I created 3 partitions in the
> >> > "Investments" cube. I understand that partitions are a way to group
> >> > data
> >> > in
> >> > a cube that is getting large. My question is how is a partition used
> >> > once
> >> > created (and processed and so forth)? the "Investments" cube has 2
> >> > partitions "Conneticut", "Vermont".
> >> >
> >> > I opened up the MDX application and can only see the 2 cubes in this
> >> > database for querying. Is it possible to query either of the
> >> > partitions
> >> > "Conneticut" or "Vermont"?
> >> >
> >> > Any explanations appreciated on the usage/purpose of cube partitions.
> >> >
> >> > Thanks,
> >> > Rich
> >> >
> >> >
> >>
> >>
> >>
>
>
>
.
- Follow-Ups:
- Re: how to use partitions once created?
- From: Akshai Mirchandani [MS]
- Re: how to use partitions once created?
- References:
- Re: how to use partitions once created?
- From: Akshai Mirchandani [MS]
- Re: how to use partitions once created?
- From: Rich
- Re: how to use partitions once created?
- From: Akshai Mirchandani [MS]
- Re: how to use partitions once created?
- Prev by Date: RE: Cube Drill through not working on some machines
- Next by Date: newbie question - browse all data MSAS cube
- Previous by thread: Re: how to use partitions once created?
- Next by thread: Re: how to use partitions once created?
- Index(es):
Relevant Pages
|