you might want to check the min and max timestamps on the returned events, I believe the response gets truncated by datetime, so you'll see that you requested events ending at 5pm but you only got events up to 3pm or whatever
it could be due to the rate limit on the export API, there's a size limit to the return object, so if you have a lot of events or they're particularly big or you're not using gzip (it doesn't look like you are, although I'm not super familiar with the language you're using) you might be getting a truncated response
We have a heavily used event property that we realized had a bug and was misleading; while we fix it, I want to go through all the reports that are affected by this property and mark them as potentially misleading. Is there any way to find these reports?
it would be helpful to have an optional flag in the API to toggle this behavior, I imagine it's rare anyone wants to export an essentially overwritten version of an event
answering my own question here: a lot of these events share the same $insert_id field, meaning that they've been "overwritten" and the frontend shows (and counts) only the most recent version of each, but the API returns all of the versions... not sure if that's on purpose or not
I'm kinda confused about the exporting API... I exported all Started Job events between these two dates, and the API returned around 900k events... however, in mixpanel, I only see 250k events... what could explain such a large discrepancy?
in any case, I did what I needed, I managed to filter the events down to something around 200k events, download them all, extract the anonymous ids and user ids from them, and send the approximately 12k identify events needed
