Extract operation about pathling HOT 4 CLOSED

Fwiw this can be simulated, perhaps, by aggregating with identifier as a grouping field. This currently appears to be a very slow operation.

from pathling.

Yes, this would be much more efficient with straight selection without any sort of aggregation. I think the implementation for this will be relatively straight forward, it is just another executor class in addition to AggregateExecutor and SearchExecutor, and a HAPI provider to wrap and expose it.

I think that the other reason that we want a dedicated operation is the need to treat the result as unbounded, so that it is practical for use with large datasets. This means making it asynchronous and writing the result somewhere for efficient retrieval, rather than in a synchronous HTTP response.

from pathling.

I suppose we could potentially optimise this case within the Aggregate operation, if it was useful...

from pathling.

I suppose we could potentially optimise this case within the Aggregate operation, if it was useful...

Well it's not exactly user friendly, but were this case optimised it would achieve the same, if you stick a .first() in each extraction query used as an aggregate.

from pathling.