Hi nickiv,
In general, you can fetch more than 1000 entities for a query by sorting and
filtering on a property - you can use the __key__ pseudo-property if you
don't need records in a particular order.
As far as constructing reports goes, your best option is probably to use the
task queue,
With an HTML page that reloads itself collect all the data that you
want in memcache.
Each reload a parameter specifies where to continue collecting.
If all collected return an HTML page that reloads to an URL that
converts the memcache content to an image, store this image in
memcache, and serve
Hi Locke,
On Tue, Sep 15, 2009 at 6:15 PM, Locke locke2...@gmail.com wrote:
I would not be returning thousands of rows to users. This would be for
background tasks executed via cron.
The hard limits appengine imposes (30 seconds, 1000 query results,
etc.) all seem quite reasonable for
use other way overcome
2009/9/17 Nick Johnson (Google) nick.john...@google.com
Hi Locke,
On Tue, Sep 15, 2009 at 6:15 PM, Locke locke2...@gmail.com wrote:
I would not be returning thousands of rows to users. This would be for
background tasks executed via cron.
The hard limits
Interesting. That seems kinda tedious.
So suppose someone were to write a wrapper around the datastore get
API that handled the 1000 entity limit by automatically splitting
queries into 1k groups, but returned them seamlessly to the user as if
they were one query. Would that be a violation of
Hi Locke,
If you want to write such a wrapper, you're quite welcome to - it's
certainly not a violation of the TOS or anything like that. We'd encourage
you to examine why you need to do this, though - do you really need to
return several thousand results in a single request? Will your users be
I would not be returning thousands of rows to users. This would be for
background tasks executed via cron.
The hard limits appengine imposes (30 seconds, 1000 query results,
etc.) all seem quite reasonable for interactive apps, but for non-
interactive background tasks, it looks like a lot of
You can sort the entities by __key__, and get the first 1000 entities.
Then add a filter for __key__ that bigger(or smaller) than the 1000th
entity, to get 1001st to 2000th entities, and so on.
2009/9/15 Locke locke2...@gmail.com:
From the documentation: maximum number of entities in a batch
