Filter looking for values based on other values

Question

Hi, is it possible to create a filter that returns items from the cache based on some criteria of other items in the cache.
E.g. suppose I have a cache containing cars. I would like to create a filter that filters out all blue cars, but only if the cache also contains a red car.
Best regards
Jan

Answer 1

Hi Jan,
it is doable in a single pass in an aggregation, but it will not be too efficient for cases where no red cars exist.
The problem is that in case of a partitioned cache, it is possible that there is a single red car in the cache on a separate node from the blue cars. In this case the blue cars should be returned. However it is also possible that there is no red car in the cache.
In this example the nodes with the blue cars from the previous example won't be able to distinguish the two cases from one another based on only their own data. Therefore they have to return the blue cars, and let the client throw away all data based on whether the existence of a red car was indicated.
If red cars are an unlikely event, it is better to first query (or aggregate for minimum traffic) for the existence of the red car, and only if it exists, then query for blue cars.
If red cars are a probable event and having as few roundtrips to the cache as possible is more important then efficiency, then it is probably better to just (in parallel) aggregate the blue cars and indicate whether a red car exists on that node or not.
For aggregations and particularly for parallel-aware aggregations, you can look at the following page:
http://wiki.tangosol.com/display/COH34UG/Provide+a+Data+Grid#ProvideaDataGrid-DataGridAggregation
Best regards,
Robert