Re: [google-appengine] Appengine datastore performance

Hi Nick, Was playing (and learning) with mapreduce but i'm starting to try to do a kind of mapreduce myself, flexible, better integrated ( http://code.google.com/p/relat/wiki/gaetips#Mapreduce), more possible maybe a bit slower. gr wim On Sat, Dec 4, 2010 at 3:25 PM, nick wrote: > no :-) > i mea

Re: [google-appengine] Appengine datastore performance

no :-) i mean: http://www.youtube.com/watch?v=AgaL6NGpkB8&feature=player_embedded minute 17:20 does that work for me? -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To post to this group, send email to google-appeng...@googlegroups.com.

Re: [google-appengine] Appengine datastore performance

This one? http://code.google.com/p/appengine-mapreduce/wiki/UserGuidePython#Doing_per-row_“reduces” On Sat, Dec 4, 2010 at 3:00 PM, Nick Heppner wrote: > what about: > > have a host entry as a parent and his logs as his child? > > i just need to write one at a time but i could read very fast ever

Re: [google-appengine] Appengine datastore performance

what about: have a host entry as a parent and his logs as his child? i just need to write one at a time but i could read very fast every entry thats a child of my parent host. does that work? nick 2010/12/4 Wim den Ouden > The splitting proces is mapreduce doing by parallel sessions for you.

Re: [google-appengine] Appengine datastore performance

The splitting proces is mapreduce doing by parallel sessions for you. There is a control module to start it from your code (normal from a dashboard), but i don't now yet if the overhead is taken to much time. There are no docs yet. In the local development server some libraries are not there, see h

Re: [google-appengine] Appengine datastore performance

ive done the mapreduce with hadoop a year ago. this is mapreduce thing has many potential but its very hard for me to understand how to use it with the datastore ;-) i that right?: i setup some "processes" that split the table and every process just parses his part and then they join the result

Re: [google-appengine] Appengine datastore performance

forgot to say, Mapreduce can allso walks (amazing fast) thru zip and text files stored in blogs and the list is growing. gr wim On Sat, Dec 4, 2010 at 1:48 PM, Wim den Ouden wrote: > This is a link to what you need for *.yaml and a very simple demo, in the > def (in this example lower_case_posts

Re: [google-appengine] Appengine datastore performance

many thanks! i'll would read me throw this stuff thanks! nick -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To post to this group, send email to google-appeng...@googlegroups.com. To unsubscribe from this group, send email to googl

Re: [google-appengine] Appengine datastore performance

This is a link to what you need for *.yaml and a very simple demo, in the def (in this example lower_case_posts(entity)) you put the code. Mapreduce runs (sometimes 20 parallel sessions) thru all entities (no filter possible yet). First a check if the entity is what you need and then your code. Map

Re: [google-appengine] Appengine datastore performance

wow. nice infos thanks! :-) if i had millions of log entries stored in one table. (l_entry: title, content, date, host) and now i want to have all log entries where host="myhost" (as a json response), how would you design the query? (this would be slow, doenst it?) (excuse my bad english :-)

Re: [google-appengine] Appengine datastore performance

http://code.google.com/p/relat/wiki/gaetips#Mapreduce gr wim On Sat, Dec 4, 2010 at 1:04 PM, nick wrote: > Hi! > > I want to handle about millions of entries in one table. Normally i would > query over every row and check for my filters (lik

[google-appengine] Appengine datastore performance

Hi! I want to handle about millions of entries in one table. Normally i would query over every row and check for my filters (like name="foobar"). Is there a best practise or some stuff to read about to get more performance? greets nick -- You received this message because you are subscrib