* caoyanjiao987 caoyanjiao987@163.com [Mon, 7 Mar 2011 20:27:01 +0800 (CST)]:
Dear Mr/Miss: Sorry to interrupt you but there are two problems to ask you for help. It puzzles us for a long time. We built a local wiki using mediawiki by downloading the page articles fromhttp://download.wikimedia.org/enwiki. However, the data cannot be imported completely by virtue of the mwdumper.jar and it stopped halfway saying that there is IllegalArguments. We tried different source data but no use. Is there something wrong with our configuration?
It's hard to work correctly with these huge dumps. Try to download a different dump. Someone would have better answer than me.
Besides, given a word pair "lion cat" as the query, theresults returned from the local wiki are differnt from that on the
web.
Wikipedia on the web returned many items for the query but the local wiki returned no results. How can make them make return the same
query
results.
Wikipedia uses Lucene search, while MediaWiki's default install search is MySQL MyISAM fulltext index - which is much simplier and slower. You may try to install Lucene http://www.mediawiki.org/wiki/Lucene
SphinxSearch is somewhat simplier alternative, probably a bit easier to install - I use it at some wikis. Dmitriy