In the pages-articles dump, you shouldn't be looking just at the content of the pages, you should be looking at the title tag. That's how you reliably find out what is the title of the page with the given content.
Alternatively, you could use page ids instead of titles (look for the id tag in the XML dump).
Petr Onderka [[en:User:Svick]]
On Fri, Oct 18, 2013 at 10:02 PM, Peyman Faratin peyman@robustlinks.com wrote:
Hi
I've a set of list page titles that i've extracted from the Category dump (where "cl_from" is of type "page")
http://www.mediawiki.org/wiki/Manual:Categorylinks_table
Now I want to extract the CONTENT of the page from the pages dump
enwiki-latest-pages-articles.xml
Although there are guidelines on how editors should mark these pages
http://en.wikipedia.org/wiki/Wikipedia:Manual_of_Style/Lists
("The titles of list articles typically begin with the type of list it is (List of, Index of, etc.), followed by the article's subject; like: List of vegetable oils.")
The majority of the times the above rule is not implemented. So my concrete question is:
- if i am consuming the pages-articles.xml dump (D) page by page, and i have
a list of pages (L) i've extracted from the category (C) links dump, then how can i check that page in pages dump file D is a member of L? The titles do not resolve the names.
For instance, if I have the page title "List of the longest Asian rivers" (http://en.wikipedia.org/wiki/List_of_the_longest_Asian_rivers) then what in that page's content (http://en.wikipedia.org/w/index.php?title=List_of_the_longest_Asian_rivers&a...) can tell me it is the same page "List of the longest Asian rivers"? None-list pages appear to place the title as first token with ''' markings.
Any suggestions of a robust solution would be much appreciated.
Best
Xmldatadumps-l mailing list Xmldatadumps-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/xmldatadumps-l