brion added a comment.

  In https://phabricator.wikimedia.org/T120452#2227240, @matmarex wrote:
  
  > As I understand these are stored as regular MediaWiki pages now, so they 
have a maximum length of 2 MB. Even naive queries pulling the whole thing into 
memory would be fast enough at these scales. If we want to think about 
performance for large data, we should first think about overcoming the length 
limitation :)
  
  
  Ah I forgot all about that. ;) That'll at least stop people from creating 
_super_ huge datasets for now... unless they break them into multiple files and 
create lua modules to stitch them back together. :) That may be acceptable 
however, and lets people prototype the kinds of crazy things they want until we 
hate it so much we decide we have to support them better.
  
  I tried generating some random tables with this PHP script: 
https://gist.github.com/brion/46469ac2df31a8eb0e179f50b1967d20
  
  I find I can only successfully save a file of under a megabyte even though 
the max is 2 megs; a 2 meg file gets rejected complaining that it's over 4 
megabytes... I notice the output that comes back to me in the edit window is 
pretty-printed, which means a lot of indentation and newlines bulking up the 
storage requirements. That might be on purpose to provide a better diff view?
  
  Seems to run reasonably fast to render locally, though it also produces a 
giant HTML table for the page view that's about 3.5 megs, which should at least 
compress ok. Editing also means resubmitting the entire data set at once with 
the edit form.
  
  A partial-edit submission system similar to section editing on wiki pages 
might be nice to reduce submission latency and bandwidth consumption editing a 
table, but that can probably wait until it needs to be paired with a 
spreadsheet-like UI.

TASK DETAIL
  https://phabricator.wikimedia.org/T120452

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Yurik, brion
Cc: Jdforrester-PERSONAL, Jdforrester-WMF, brion, ThurnerRupert, intracer, 
TerraCodes, Pokefan95, gerritbot, -jem-, Bawolff, MZMcBride, Alkamid, 
Milimetric, Thryduulf, JEumerus, MarkTraceur, Yurik, Matanya, ekkis, matmarex, 
Lydia_Pintscher, Aklapper, Steinsplitter, StudiesWorld, DannyH, Riley_Huntley, 
D3r1ck01, Izno, JAllemandou, Wikidata-bugs, aude, El_Grafo, Ricordisamoa, 
Shizhao, Fabrice_Florin, Mbch331, Jay8g, Krenair, jeremyb



_______________________________________________
Wikidata-bugs mailing list
Wikidata-bugs@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs

Reply via email to