If Drizzle is for Clouds, then we ought to think some about what Cloud use-cases are really common and really hard to do with standard SQL (and especially with MySQL).
Here's one: any three of group by, order by, limit, and offset. Here's another: queries to show results in some ranked order. (OK, I'm just sneaking in a double vote for the one above). How can we make these easier to write and faster to perform? Most of the web aps I get asked to look at have trouble with these. In some cases these types of queries can even be the overwhelming majority of the load. Not that these are things to build now -- just wanted to throw these common-yet-hard things out there, and propose the idea that some non-relational things can help a lot here. Sphinx does a great job at this -- it totally slays where MySQL totally falls down for queries like "get the most important X by Y criterion and give me results Z through W, but also tell me how many there are in total." It does this in part because it doesn't require you to do relational things like joining to a subquery in the FROM clause. What other things are hard for MySQL now, and are just kinda hard to do purely because relational-ism makes it so? Some places where MySQL has broken relational-ism, like ON DUPLICATE KEY UPDATE, have really been beneficial -- more? How about inserting and getting the result back at the same time? (Postgres does this already) Or inserting into two tables at once? Or deleting from one table and inserting into another at the same time? Or deleting while getting back the deleted rows? (That would be a "queue", but it would be useful for a lot more than that -- think data archiving -- DELECT FROM foo INTO foo_archive WHERE ....") _______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
