Results 1 to 10 of 10
  1. #1
    Join Date
    Jun 2007
    Location
    Dresden, Germany
    Posts
    31

    Data Warehouse or not???

    Hi @all!

    I have 5 different data sources (relational databases, flat files, ...). I extracted the relevant data from them, transformed them (clean, combine, standardize, ...) and combined them in one database with different tables. Looks like a normal ETL process, right? Now I have to write a programm which does some analyses with this new database (mostly select, which generates different pivot tables).

    My question now is fundamental: Can say that this new database is a data warehouse, even without dimensional tables in it? I know that a DWH normally consists of dimensional tables ready for BI apps, but at the moment I do only 2D queries (pivot tables).

    Thx in advance!
    sp00ky

  2. #2
    Join Date
    Jun 2003
    Location
    Ohio
    Posts
    12,592
    I'd say it is more likely to be a Data Warehouse BECAUSE it does not contain "dimensional tables".
    Dimensions, star schemas, etc, are suitable for Data Marts. They are woefully inadequate for Data Warehouses. Kimball and his ilk sell them as "Data Warehouses", simply because they are an easy product to market. But they fall far short of the promise of a true data warehouse.
    If you want to know what a data warehouse is truly supposed to be, read Inmon, not Kimball.
    And by my definition, you have a data warehouse. Multiple data sources, multiple business disciplines, related in a single (hopefully normalized) schema.
    If it's not practically useful, then it's practically useless.

    blindman
    www.chess.com: "sqlblindman"
    www.LobsterShot.blogspot.com

  3. #3
    Join Date
    Jun 2007
    Location
    Dresden, Germany
    Posts
    31
    Yes you are right, the last books I read were the books of Kimball! The Inmon 4th Edition I have also here... So you think, a DWH must not contain dimensional tables, it can be LIKE a normal database which is combined from n databases?! And Data marts consist of dimensional tables, star scheme etc, right?

    I need this for my diploma thesis: Can I say that the DWH term is not clear defined and Kimbal uses the DataMart term from Inmon as DWH and defines a new DataMart term, which is in fact an extended View of a DataMart?!?!

  4. #4
    Join Date
    Jun 2003
    Location
    Ohio
    Posts
    12,592
    Kimball does not promote Data Warehouses. He promotes seminars and books. The notion that a simple one-size-fits-all star schema will fulfill any customer's data warehouse needs is as easy to sell as a cure-all snake-oil elixir.

    His methods are appropriate for datamarts, which are narrowly focused to a single business area.

    Inmon's philosophy is to have a single date warehouse consolidating multiple business disciplines. If necessary, data marts can be generated from the data warehouse.

    Kimballs philosophy is simply to create many data marts (star schemas). The collection of these data marts is what he considers a datawarehouse. But of course, this in no way fulfills one of the goals of a data warehouse which is to have a single record of authority for data. To say nothing of the administrative headache that inevitably results.
    If it's not practically useful, then it's practically useless.

    blindman
    www.chess.com: "sqlblindman"
    www.LobsterShot.blogspot.com

  5. #5
    Join Date
    Dec 2007
    Location
    London, UK
    Posts
    741
    Quote Originally Posted by sp00ky
    So you think, a DWH must not contain dimensional tables, it can be LIKE a normal database which is combined from n databases?!
    Absolutely. A data warehouse is best implemented as a Normal Form data model. Its key features are that it is subject oriented, non-volatile and consistent.

  6. #6
    Join Date
    Jun 2007
    Location
    Dresden, Germany
    Posts
    31
    Ok, thx for ur comments!

  7. #7
    Join Date
    Jun 2003
    Location
    Ohio
    Posts
    12,592
    Quote Originally Posted by sp00ky
    So you think, a DWH must not contain dimensional tables, it can be LIKE a normal database which is combined from n databases?!
    A database, normalized, should ideally be a MORE accurate representation of the actual business model, since it is not constrained or hampered by the requirements of supporting OLTP.
    Model your business, across multiple disciplines, and you have a data warehouse design.
    If it's not practically useful, then it's practically useless.

    blindman
    www.chess.com: "sqlblindman"
    www.LobsterShot.blogspot.com

  8. #8
    Join Date
    Jun 2007
    Location
    Dresden, Germany
    Posts
    31
    In fact, I have no business data in my data warehouse. I work for an automotive concern and I'm analyzing crash data in correlation with spare parts usage. The data I use, comes from 5 different databases/snapshots. I have designed a custom ETL process and have now an atomic and normalized database. After all, I think I can "sell" it as DWH!

  9. #9
    Join Date
    Feb 2004
    Location
    In front of the computer
    Posts
    15,579
    Quote Originally Posted by sp00ky
    In fact, I have no business data in my data warehouse.
    This could easily qualify as the "Quote of the year" for me.

    Based on this statement, you don't realize that your data truly IS business data!!! It might not be invoices or inventory, but your business needs this data to do business, and in my mind that need makes it business data. At least at the moment, I can't think of a any better definition for "business data" than that.

    -PatP
    In theory, theory and practice are identical. In practice, theory and practice are unrelated.

  10. #10
    Join Date
    Jun 2007
    Location
    Dresden, Germany
    Posts
    31
    Quote Originally Posted by Pat Phelan
    This could easily qualify as the "Quote of the year" for me.
    LOL... thx for the direct words

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •