hello 1 importing data through an application file can couse outofmemory exceptions: crank up the memory settings of the vm (good idear any way). The method is not for regualar imports (adding data) i think, only the first batch
2 we are using the xml importer regularly, and it works ok. We had some trouble to get it to work, but now it works ok. there was one snag: the <create> tag has a timeout attribute that we had to learn about the hard way (not documented). it takes a value in seconds. Make that a nice high number, those imports can take a long time (hours). By writing your own mergers you have a lot of control over the merge process. If you want i can send you exaples of an import file and Merger and Finder classes. If for some reason the import goes wrong the transaction is lost and you can start again. No data will be imported at all. > -----Oorspronkelijk bericht----- > Van: Ignacio Renuncio [mailto:[EMAIL PROTECTED] > Verzonden: woensdag 2 juni 2004 17:19 > Aan: [EMAIL PROTECTED] > Onderwerp: Experiences importing large amounts of data > > > > Hi again, > > I was thinking about the possibility of importing moderately > large amounts > of data into MMBase, several thousands of records. > > Is the XML importer the option to use? The recommended one? > > Has anybody tried to import directly through SQL? Is it possible? What > pros/cons does it have? > > > > Thanks a million again, > > > Ignacio.- > >
