(HAPI-FHIR 6.6.0, JPA)
In order to just 'count' the stored resources for summarizing purposes, seems that the
/<Resource>?_count=0 operation (or
summary=count) are very slow when number of resources are starting to be big enough....
In paralell we have the
/$get-resource-counts operation, that is low level, effective and fast, BUT it includes the resources marked as deleted (functional delete, entry remains) and is a bit confusing and creates misunderstandings.
I am considering to create a paralell custom operation, or extend the current, to avoid included such deleted ones in the count, but I am surprised that it is not something already raised, since the conformance counters do not exclude eliminated resources either.
Is it because there is some alternative that I am overlooking?
Thanks in advance