Results 1 to 6 of 6
  1. #1
    Join Date
    Jun 2010
    Posts
    81

    Unanswered: Important Datawarehouse Criteria

    In designing and building a datawarehouse database - what are the 3 main points or criteria.

    In a reporting database I would have thought this to be disks for fast access and the network.
    Also to take into consideration for backup performance speeds and ETL

    Can someone clarify please

  2. #2
    Join Date
    May 2010
    Posts
    56
    Denormalized, less indexes and Transaction Recovery is not necessary as in databases transactions.

  3. #3
    Join Date
    Dec 2007
    Location
    London, UK
    Posts
    741
    I don't believe you can summarise the "main" aspects of such a broad subject with just three points. I'd say that logical design and data integrity features ought to come at or near the top if the list.

    With all respect to sudhirc212, as a general rule I would disagree with the first of his points and I would question the assumptions behind the latter two. Boyce-Codd / 5th Normal Form is typically the best model for a data warehouse and ought to be the starting assumption in my view. In a warehouse, indexes remain very important from a query optimisation point of view. Transaction recovery certainly can be important depending on how the warehouse gets updated and the frequency and volume of updates.
    Last edited by dportas; 06-22-10 at 15:34.

  4. #4
    Join Date
    Aug 2003
    Location
    Where the Surf Meets the Turf @Del Mar, CA
    Posts
    7,776
    Provided Answers: 1
    >what are the 3 main points or criteria.

    1) data loads quickly

    2) reports run quickly

    3) reports are accurate
    You can lead some folks to knowledge, but you can not make them think.
    The average person thinks he's above average!
    For most folks, they don't know, what they don't know.
    Good judgement comes from experience. Experience comes from bad judgement.

  5. #5
    Join Date
    Jun 2010
    Posts
    81
    Thanks for your comments. However what init parameters would govern the database to ensure data is loaded quickly and reports are run quickly? Another noticeable point from a DBA point of view is backups. Due to the size of any DW DB I would have thought the types of disks, which the data resides on would be an important factor in not only I/O performance but to also ensure any RMAN backup was performed to its optimum performance.

  6. #6
    Join Date
    Jun 2003
    Location
    Ohio
    Posts
    12,592
    Provided Answers: 1
    Quote Originally Posted by sudhirc212 View Post
    Denormalized, less indexes and Transaction Recovery is not necessary as in databases transactions.
    The incorrect assumption that data warehouses should be denormalized has been repeated so often now that it has attained an almost urban-folkloric aura of authenticity. We can thank Kimball for that, I suppose, but it really deserves its own page on snopes.com.

    In my opinion, if you start with a denormalized schema then you do NOT have a data warehouse. You have, at best, a reporting database.
    If it's not practically useful, then it's practically useless.

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

Posting Permissions

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