• (nodebb)

    "I think they did that for performance."

    Performance art I hope?

  • LZ79LRU (unregistered)

    Can I play with madness. - Yes, go into DB design.

  • Sauron (unregistered) in reply to LZ79LRU

    A linked list of database? WTF?

    The guy who created that is not a clown, he's an entire circus.

  • Confucius (unregistered)

    Relational Databases are the closest we come to actual Wizardry in the modern world. Only the Initiated know how they work. Sometimes you have to take arcane actions to get your result. They frequently work in mysterious ways.

  • Randal L. Schwartz (github)

    Ahh yes, the π-normal-form.

  • (nodebb)

    Storing the date of events in the name of the table? That's not going to end well, ever.

    I'm sure that if you describe it as that, at least some folks are going to have first thoughts about it. (I would say "second thoughts", but that would imply that those people had given it even a microsecond of thought in the first place.)

  • (nodebb)

    The real WTF is not using ISO dates in the table names.

  • (nodebb)

    Someone reinvented database-partitioning. Very badly of course, because Raisins.

  • (nodebb)

    Seems like the designers of this scheme reinvented database partitioning.

  • (nodebb)

    To those (justifiably) thinking this is a clownish version of partitioning - well not really, because according to the article, there are materialized views which "join" (did they mean "unioned"?) all of the tables together, which undoes the benefit of partitioning, and furthermore, makes the whole thing that much worse.

  • (nodebb) in reply to Mr. TA

    None of the people you seem to be addressing said or even implied that the partitioning had been reinvented well. Approximately 58.27% of TDWTF articles come from someone's attempt to reinvent a standard tool or technique.

  • Mark (unregistered)

    I suspect this was created before partitioning was available & migrated ever since. Not that uncommon in the days before partitioning.

  • (nodebb)

    One of the more difficult things to learn when designing a relational database is learning when not to "do things my way" when the database engine is already far more capable, fast and efficient at solving that "performance problem" than you are.

  • WTFGuy (unregistered)

    Something I say a lot in other contexts is "You must be smarter than your tools". That gets harder and harder as our tools get smarter. TDWTF proves a lot of people are not nearly as smart as their tools.

  • (nodebb) in reply to staticsan

    One of the more difficult things to learn when designing a relational database is learning when not to "do things my way" when the database engine is already far more capable, fast and efficient at solving that "performance problem" than you are.

    And, only after that, learning when and how you should do things differently because you do sometimes know better than the database engine, so you can steer it away from a poor access path and towards a better one. You can't do it well until you understand your data, your indexes, the general behaviour of your database's optimiser, and how to read a query plan.

Leave a comment on “Table This Until Tomorrow”

Log In or post as a guest

Replying to comment #:

Log Out

« Return to Article