Cube summary: In an MV implementation, a cube might be a file with a multipart key and each part of the key being a foreign key to another file. Additionally, instead of virtual fields for summing such numbers as total-sales-per-salesperson-per-quarter, the cube might actually store these sums just in case someone wants to see them then they do their online analytical processing (OLAP), for example.
The most important aspects of a cube are the FACT table (that's the one with the multipart key), the DIMENSION tables (the ones the foreign keys point to) and the MEASURES (the values to be summed, counted, etc when slicing and dicing the data). RDBMS users often rehost their data in a data mart or warehouse using a "star schema" (the name for a fact and dimension table design) so they can report against the data. [And as an aside, PICK folks sometimes also port their data to a relational database so they can then make stars and put in an OLAP cube so they can then report against it as if it were multivalued data. We end up doing that due to a lack of standard reporting solutions other than the character-based MV query tools. That's why I promote such tools as Informer at www.entrinsik.com -- we could bypass a lot of steps if we don't buy into the relational mistakes -- couldn't resist the soapbox opportunity, sorry]. --dawn Dawn M. Wolthuis Tincat Group, Inc. www.tincat-group.com Take and give some delight today. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Mark Johnson Sent: Sunday, April 18, 2004 2:49 PM To: U2 Users Discussion List Subject: Re: GUI from Mv code Re: Crystal Reports Not to be out of touch, but what is a Sales Cube. I saw Swordfish and I hope that itn't it. Thanks. ----- Original Message ----- From: <[EMAIL PROTECTED]> To: "U2 Users Discussion List" <[EMAIL PROTECTED]> Sent: Sunday, April 18, 2004 3:20 PM Subject: Re: GUI from Mv code Re: Crystal Reports > In a message dated 4/18/2004 10:18:11 AM Eastern Daylight Time, [EMAIL PROTECTED] writes: > > > For example, aligning this back to the original post, rather than attempting to use CR for reporting, I'd simply create a Vis�ge.BIT cube to give the users free-form enquiry and data exploration facilities into their UV database. In our case we have replaced >300 sales analysis reports currently provided in our R5 system with a single Sales Cube - and this is one of the facilities that makes Vis�ge "better" than > > AccuTerm ! > > Warning! Salesman quote! > C'mon Ross :) A dataset that includes 50 fields and I only want to see 6 on my report. So you replace that report with a cube where I (the user) have to figure out exactly what I want to see, build the proper query statement, format statement, display statement etc and then figure out how to tell the system to remember my statement so next time I don't have to THAT all again. > > You replaced the 300 reports, with one huge cube where you STILL have users recreating (or trying to) their original 300 reports... > Users don't want to see 50 fields, they want to see the 6 fields they've been analyzing for the past 3 years ... > Will > -- > u2-users mailing list > [EMAIL PROTECTED] > http://www.oliver.com/mailman/listinfo/u2-users -- u2-users mailing list [EMAIL PROTECTED] http://www.oliver.com/mailman/listinfo/u2-users -- u2-users mailing list [EMAIL PROTECTED] http://www.oliver.com/mailman/listinfo/u2-users
