On 12/21/2007 05:02 PM, Lars Aronsson wrote:

Hi Johann,

...

When I just want current pages with no discussion and no history but with all the templates and category lists working correctly, which of all the files in http://download.wikimedia.org/enwiki/latest/

That should be the pages-articles.xml "the one most people will want".

What puzzles me about that is that there are so many other files in the download dir ... don't I need any of these?

When I last tried that (I downloaded pages-articles.xml, converted with Xml2sql and imported with mysqlimport), templates did not show properly and no category lists where available at all.

...

do I have to download and how do I have to import the data so that the import is complete and fast?

Fast? Haha, no, it won't be.

You should try it out on a smaller language than English. Try Faroese (fo) with 2,700 articles or Anglosaxon (ang) with 900 articles.

Well, I meant *relatively fast* :) It seems there are different ways to do the import which might be differently fast, given the same dump.

I also wonder if there are any settings that would speed up the import into the local DB ... e.g., I turned off the binary log of Mysql which seemed to speed up things. I only want to read the local copy, not update anything.

Cheers, Johann