Bulk export performance experimentation

[Michael Terry]

Hello folks! Does anyone have any tips or tricks for improving the performance of bulk exports?

We just started playing with bulk exporting at Boston Children's Hospital, against a test Group of 4k patients, and are finding that limiting the scope with _since doesn't really seem to take less time.

For example, exporting all Encounters from this Group in the past couple days (zero) takes a bit over an hour. And exporting all Encounters in the last month (1.5k) takes the same amount of time.

Has anyone found a clever _typeFilter parameter or similar trick that does improve the query performance? Maybe it's not possible based on how it's implemented, but thought I'd ask.