RE: [PEAR] MDB 2.0 release and features.

> From: Fabrizio Balliano [mailto:fabrizio.balliano@crealabs.it] > Sent: Friday, August 08, 2003 7:04 PM > We started developing it with PEAR::DB, now we've switched to PEAR::MDB > because we see many possible new features with the MDB Manager. > > I'm trying to autodetect database field type using PEAR::MDB, but we see > that: > > - tableInfo does not return PEAR::MDB field types but the native types. Yes, tableinfo is basically just a copy paste from PEAR::DB > - listTableFields needs a table and does a query for every call. > - getTableFieldDefinition needs a table and does a query for every call. These methods have been written for reverse engineering. So they are the closest to what you are looking for. However they only really exist for mysql atm and partially for pgsql. But they were never meant to be used at run time as they have an insane amount of overhead. > Is it possible to know the PEAR::MDB compatible field type having a > result set and not a table? > Eg: $rs = $mdb->query('SELECT * FROM tab_1 JOIN etc... etc..'); > $real_data_types = $mdb->tableInfo($rs) > $mdb_data_types = $mdb->tableInfoMDBTypes($rs) > Could this be implemented in the 2.0 version? The philisophie that MDB follows is that all type related information needs to be passed to MDB by the programmer and I do not see this changing in the future. But there is hope for you yet! I am collaborating with Alan Knowles to build a DataObject class on top of MDB. This class will use the information from the XML schema file. To increase performance the parsed schema will be stored in a serialized format that is quicker to parse. > Are there some news about the 2.0 release date? Well I would say it will take another 1-2 months for MDB 2.0 to be finished. I am mostly done with the refactoring, but there are still a few things that need to be thought out. MDB_DataObject will get finished sometime this fall. Unfortunately I don't have more time for this project. Donations are welcome to speed up the process ;-) > Excuse me for all these questions. No problem at all. FYI: I am actually separating the Manager from MDB. The new manager will be insanely more powerful and flexible. However to take advantage of the datatype abstraction at runtime you still need MDB. The new Manager will probably be finished in 2-3 months. Regards, Lukas