A few days ago I read an excellent post by Martin Elwin on creating a query cache using Coherence. To summarize, he uses a Coherence Filter (an object that contains query criteria) as a key and the results of the query (in this case, a set of keys for …

Thanks for your comments. There is some overlapping functionality between this and CQC. Here are the ways in which they differ:

1. CQC will register a map listener internally, thus ensuring the client gets the update right away, whereas the query cache will be stale for a period of time.

2. CQC only accepts one filter at a time, the query cache can be used to cache many filters.

The CQC is a good choice for clients that need real time access to a subset of data (defined by a single filter), whereas the query cache is a good option for caching commonly executed queries.

Thanks,
Patrick

Just wondering how is this Continious query cache(CQC).
Refernces: http://wiki.tangosol.com/display/COH32UG/Continuous+Query

Please let me know your thoughts on this.

Regards
Vikas

We have several use cases where we want to get parent objects based on the child object’s properties. For example, get all parent objects that have a child object with value X. Given a model where the parent has a collection of child objects and the child objects have a reference to the parent (parent.getChildren(), child.getParent()) and assuming that the parent objects and child objects are stored in separate named caches, our initial reaction was to construct a filter:

Filter filter = new EqualsFilter(“getName”, “foo”);
NamedCache cache = CacheFactory.getCache(“childCache”);
cache.addIndex(new ChainedExtractor(“getName”);
Aggregate parentAg = new DistinctValues(new ChainedExtractor(“getParent”));
Set parents = cache.aggregate(filter, parentAg);

We found it was much more performant to change the search to be more like so:

Filter filter = new EqualsFilter(“getName”, “foo”);
NamedCache cache = CacheFactory.getCache(“childCache”);
cache.addIndex(new ChainedExtractor(“getName”);
cache.addIndex(new ChainedExtractor(“getParent.getId”);
Aggregate parentAg = new DistinctValues(new ChainedExtractor(“getParent.getId”));
Set parentKeys = cache.aggregate(filter, parentAg);
Map parents = CacheFactory.getCache(“parentCache”).getAll(parentKeys)

We have created prebuild responses and have been looking at doing a query cache as you suggested. There seems to be a lot of patterns out there, such as these, that could become part of the coherence incubator project.

aloha
Paddy

We can certainly consider putting this in the Commons project. Is the above different from your implementation?

Thanks,
Patrick

-paddy

/M