Your best bet would probably be to send a query for one record, get the
number of matching records, and then repeat several requests for a
single record starting at a random number from 1 to the number of
records discovered on the first query.
So if you do a search for state:NJ with start=0 and rows=1, you can pull
out the /response/result[numFound] value. Next, you can query it for
the same search criteria with a start={random number from 0 to
numFound-1} and rows=1, grabbing the result each time. The first set of
search results will likely still be in the cache for you to get 10 such
queries through quickly. The trick is to make sure you don't pick the
same random number twice. Perhaps an associative array using the number
as the key to you can check the existence of.
Ryan Grange, IT Manager
DollarDays International, LLC
[EMAIL PROTECTED]
480-922-8155 x106
Evgeniy Strokin wrote:
I want to get sample from my search result. Not first 10 but 10 random (really
random, not pseudo random) documents.
For example if I run simple query like STATE:NJ no order by any field, just the query and get 10 first documents from my result set, will it be random 10 or pseudo random, like first 10 indexed or something like this?
Thank you
Gene