8 events
when toggle format what by license comment
Aug 24 '14 at 13:42 comment added Gili I did, and got surprising results: dba.stackexchange.com/q/74693/4719 I admit this isn't a typical example of normalization, but it doesn't highlight that JOINs are (still) very expensive.
Aug 23 '14 at 12:56 comment added Joel Brown @Gili - re: the cost of joins: There is no right answer to that question aside from "it depends". Join cost is impacted by many factors. Whether they are negligible is even harder to answer, because that is ultimately subjective. The best way to answer you question is to mock up some test data and do volume testing. Try it both ways and see if you can tell the difference using real world data volumes (whatever that implies for your application).
Aug 22 '14 at 14:38 comment added Gili I want to segregate along commonly used columns for design reasons (separation of concerns, increased code reuse). Has anyone posted an estimate of the cost of such joins? Are they negligible or something I should worry about long-term?
Aug 22 '14 at 12:03 comment added Joel Brown @Gili - If your assumption were true then this case wouldn't apply. Segregating large and infrequently needed columns allows more rows to fit on a page, thereby allowing faster retrieval of the commonly used columns. Obviously reading the segregated columns along with the commonly used columns would be slower since a join is necessary.
Aug 21 '14 at 12:37 comment added Gili You have some subset of columns that are very wide and you want to segregate them physically in your storage for performance reasons. How does segregating them improve performance (assuming the columns are always accessed every time the main table is)?
Aug 11 '12 at 12:11 history edited Joel Brown CC BY-SA 3.0
Added another example (for differentiated security reasons)
Mar 22 '12 at 20:35 history edited Joel Brown CC BY-SA 3.0
Added additional information about the motivations for using 1:1 relationships.
Mar 22 '12 at 13:57 history answered Joel Brown CC BY-SA 3.0