The attribute name is batchSize="-1" (it is case sensitive) . Tjis ensures that Mysql driver fetcches row by row http://wiki.apache.org/solr/DataImportHandlerFaq
On Mon, Nov 3, 2008 at 9:17 PM, sunnyfr <[EMAIL PROTECTED]> wrote: > > Hi Shalin, > * > I would like to know if you just used batchsize = -1. > When I do that I use all Mysql's memory and it's a problem for the database > and other process on it like update insert ... > It will keep my database busy for 10hours, it's too much, is there a way to > manage it differently ? > > Thanks a lot, > > > > Shalin Shekhar Mangar wrote: >> >> Hi Rusli, >> >> Without going into specifics, I can tell you that we're using >> DataImportHandler successfully in production with many millions of >> documents >> being imported from databases. Try it out and let us know if you face any >> problems. DataImportHandler will be part of Solr 1.3 (the next release) >> and >> may be committed quite soon after which you may not need to fiddle with >> patches. >> >> Disclaimer: Since I'm one of the developers, I may be biased :) >> >> On Tue, Jun 10, 2008 at 6:11 PM, Rusli Ruslakall >> <[EMAIL PROTECTED]> >> wrote: >> >>> I wanted to see what people think about using DataImportHandler and >>> importing directly from mysql rather than with XML. >>> >>> I have already programmed mysql -> XML in PHP and configured Solr to >>> import it when I realized I could use DataImportHandler and skip the >>> XML stage. Are there any negatives of using DataImportHandler? I would >>> have to use the patch that complicates the process I guess when its >>> kind of tempting to try to use Solr as much vanilla as possible. >>> >>> What is the general method when using database information? >>> >> >> >> >> -- >> Regards, >> Shalin Shekhar Mangar. >> >> > > -- > View this message in context: > http://www.nabble.com/XML-vs-mysql-import-with-DataImportHandler-tp17754471p20305117.html > Sent from the Solr - User mailing list archive at Nabble.com. > > -- --Noble Paul