Filter API will hit the database if not found in Cache???

Question

I wrote a sample program using Filter API i.e. LikeFilter, EqualsFilter which will query the cache (i.e. Map in memory) based on the criteria. But what if it doesn't found the entries in Cache . Will it then query the database with similar criteria.
I don't think so ... Please correct me if i am wrong.
I need to implement Pre-loading the cache and the link http://download.oracle.com/docs/cd/E14526_01/coh.350/e14509/preloadcache.htm#CACCFCFJ
use the Filter API. I am not sure if i can rely on pre-loading the cache from database.
Any sample example preloading the data from database into cache will be greatly appreciated
Regards,
Bansi

Answer 1

Then you won't need the Invokable code (anyway its owned by my employer ;-)
All you need to do is write a simple program something like this :
1. Query everything from the database "select * from table"
2. Iterate through the ResultSet converting them all into objects
3. Cache.put() these objects into Coherence.
Now some optimisations. Re :
1. You have a single query going to the database and will be connecting to a single cluster member. You could divide the query up by some logical division (eg. Customer / Date etc.) and then have many clients putting in concurrently, but it raises the complexity.
2. It is also possible to store List tuples in Coherence, I once worked on a very nice project along those lines.
3. Cache.putAll() is much faster than put() and you should batch these into sizes of 1000 for maximum performance.
However, I would ignore these optimisations for now (tho 3. is the most useful).
You will have to manually kick off this process when you restart your cluster (or put it in your startup script). People generally write a cluster restart script. Tho JavaServiceWrapper is very nice for restarting nodes that run out of memory. When you get really complicated you can use something like FabricServer to dynamically control your cluster, but lets get the basics working now.
Best, Andrew.