William, Good list so far. I think you've got most of them.
Procedures are a big missing item right now, especially in the oracle world I spend so much time in. Oracle has that all-in-the-db philosophy down to a science. However, I believe the next version of MySQL includes stored procedures. In fact, I think the proc language is perl. One other thing that would be important, and I think CDBI has it, is some sort of method to edit data before and after it goes into the database. MySQL's timestamps, or oracle's date fields might need to be translated into something usable in your application. (So wrap the column value in a TO_DATE or TO_CHAR in oracle, or do the work in perl. Faster the first way, of course.) I'll think about it, seems like there are a couple of other things that should matter too, but it is early yet, my brain needs more coffee. The poop page looks very interesting indeed. Good compare/contrast for those modules. I will investigate that more thoroughly later on. thanks, caleb On Mon, 4 Oct 2004 11:39:59 -0400, William McKee <[EMAIL PROTECTED]> wrote: > Hi Caleb, > > You bring up several important features that a decent database > abstraction layer should handle. Perhaps it would be useful to draw up a > table comparing/contrasting some of the existing modules. > > Here's a take on the list based on the recent posts: > > * Joins - support for joins > > * named sequences - (CDBI has support for this; I use it regularly > with PostgreSQL) > > * procedure calls - I don't use these but don't think that CDBI has > direct support for them; custom sql would need to be employed > > * custom sql - easily call custom sql to perform non-trivial queries > > * customizable where clauses - use an existing table definition to > query the table with a specific where clause (CDBI supports this > ability via the retrieve_from_sql function > > * Performance - certainly important, but this figure will vary based > on the system, the database backend and the module; I think having a > benchmark suite that could be run on the target system would prove > most useful (in case anyone has some extra tuits lying around) > > Perhaps this would be useful on the wiki. There is already a comparison > of several frameworks on the poop page (Perl object-oriented > persistence)[1] as well. > > William > > [1] http://poop.sourceforge.net/ > > > > -- > Knowmad Services Inc. > http://www.knowmad.com > > --------------------------------------------------------------------- > Web Archive: http://www.mail-archive.com/[EMAIL PROTECTED]/ > http://marc.theaimsgroup.com/?l=cgiapp&r=1&w=2 > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > -- caleb http://www.ruejulesverne.com --------------------------------------------------------------------- Web Archive: http://www.mail-archive.com/[EMAIL PROTECTED]/ http://marc.theaimsgroup.com/?l=cgiapp&r=1&w=2 To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
