28 Nov
2008
28 Nov
'08
12:14 p.m.
On 28.11.2008 12:56 Uhr, Chris Withers wrote:
Andreas Jung wrote:
There is a difference pulling 500k brains out-of-the catalog compared to traversing a ZODB with 500k object. I assume that the catalog solution is slightly faster :)
Yes, well, you kno what they say about assumptions...
ok, you want the hard facts :-) CMF Site with subfolder containing 87k objects: Obtaining the brains from the catalog: 25 seconds (uncached, basically only ZODB load operations) Same with the ZCatalog within the ZODB cache: <1 milli second Using ZopeFind: I stopped after 10 minutes... ANdreas