![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
Hello there everyone Can someone answer what may appear to be a stupid question please ? If I install analysis Services 2005 can I use my exsisting SQL 2000 Datawarehouse to supply the dimension tables ? |
#3
| |||
| |||
|
|
Yes, it's possible. Ohjoo "jimmy77" <jimmy77 (AT) discussions (DOT) microsoft.com> wrote in message news:61CF94B6-56DA-4BCC-84E4-A8D71B3724B2 (AT) microsoft (DOT) com... Hello there everyone Can someone answer what may appear to be a stupid question please ? If I install analysis Services 2005 can I use my exsisting SQL 2000 Datawarehouse to supply the dimension tables ? |
#4
| |||
| |||
|
|
There are things tricky to understand. Suppose the time dimension table whose columns are year and quarter. Year Quarter ---------------- 2005 Q1 2005 Q1 2005 Q1 : 2006 Q1 2006 Q1 2006 Q1 In AS 2000, if the time dimension hierarchy is Year->Quarter, each quarter member under year member is distinguishable. In AS 2005, each attribute hierarchy is according to each column of dimension table. Here we can think Year and Quarter attribute hierarchy respectively. Each attribute hierarchy is flat consisting of just two levels (all level and bottom level). The members at bottom level are unique and come from the according column using distinct query. For example, Q1, Q2, Q3 and Q4 for Quarter attribute hierarchy. Here you need to understand that Q1 is Q1 itself and does not belong to any year. Exactly, Q1 is the same as (Q1, [All Year]). In other words, if you look at some measure for Q1, it means aggregate from all Q1s of every year. If you want to look at Q1 for 2005, you also have to select [2005] from Year attribute hierarchy. And it means (Q1, [2005]) Now you can consider having all quarter data distinguishable such as 05Q1, 05Q2, 05Q3, ..., and 06Q4. If then, [05Q1] itself means Q1 for 2005. In other words, ([05Q1], [All Year]) or ([05Q1], [2005]) Which way is better? It's up to you to choose which way. Whichever way you choose, you can finally make every combination of years and quarters. I hope this helps you. Ohjoo "Ohjoo Kwon" <ojkwon (AT) olap (DOT) co.kr> wrote in message news:%23xLuEMABHHA.3316 (AT) TK2MSFTNGP02 (DOT) phx.gbl... Yes, it's possible. Ohjoo "jimmy77" <jimmy77 (AT) discussions (DOT) microsoft.com> wrote in message news:61CF94B6-56DA-4BCC-84E4-A8D71B3724B2 (AT) microsoft (DOT) com... Hello there everyone Can someone answer what may appear to be a stupid question please ? If I install analysis Services 2005 can I use my exsisting SQL 2000 Datawarehouse to supply the dimension tables ? |
#5
| |||
| |||
|
|
That is very helpful Ohjoo thank you very much, my next question I am going to ask though is can you installl just the AS2005 part without installing or purchasing the whole sql2005 and if so would you recomend this in you honest opinion as there is a shortage of staff were I work and I want to just develop cubes without the need for the whole datawarehouse migration aswell. sorry to be a pain but any resources or articles you could provide on this would be sincerely appreciatted , thank you again. "Ohjoo Kwon" wrote: There are things tricky to understand. Suppose the time dimension table whose columns are year and quarter. Year Quarter ---------------- 2005 Q1 2005 Q1 2005 Q1 : 2006 Q1 2006 Q1 2006 Q1 In AS 2000, if the time dimension hierarchy is Year->Quarter, each quarter member under year member is distinguishable. In AS 2005, each attribute hierarchy is according to each column of dimension table. Here we can think Year and Quarter attribute hierarchy respectively. Each attribute hierarchy is flat consisting of just two levels (all level and bottom level). The members at bottom level are unique and come from the according column using distinct query. For example, Q1, Q2, Q3 and Q4 for Quarter attribute hierarchy. Here you need to understand that Q1 is Q1 itself and does not belong to any year. Exactly, Q1 is the same as (Q1, [All Year]). In other words, if you look at some measure for Q1, it means aggregate from all Q1s of every year. If you want to look at Q1 for 2005, you also have to select [2005] from Year attribute hierarchy. And it means (Q1, [2005]) Now you can consider having all quarter data distinguishable such as 05Q1, 05Q2, 05Q3, ..., and 06Q4. If then, [05Q1] itself means Q1 for 2005. In other words, ([05Q1], [All Year]) or ([05Q1], [2005]) Which way is better? It's up to you to choose which way. Whichever way you choose, you can finally make every combination of years and quarters. I hope this helps you. Ohjoo "Ohjoo Kwon" <ojkwon (AT) olap (DOT) co.kr> wrote in message news:%23xLuEMABHHA.3316 (AT) TK2MSFTNGP02 (DOT) phx.gbl... Yes, it's possible. Ohjoo "jimmy77" <jimmy77 (AT) discussions (DOT) microsoft.com> wrote in message news:61CF94B6-56DA-4BCC-84E4-A8D71B3724B2 (AT) microsoft (DOT) com... Hello there everyone Can someone answer what may appear to be a stupid question please ? If I install analysis Services 2005 can I use my exsisting SQL 2000 Datawarehouse to supply the dimension tables ? |
![]() |
| Thread Tools | |
| Display Modes | |
| |