Querying Date Ranges

EventStoreDB
#1

Hi,

I’m working on using EventStore for a logging implementation and I am trying to come up with a way to query events within a date range.

I want to be able to provide a start and end date range and then pull all events stored between that range.

I have done some digging but can’t seem to find much guidance.

Does any one have any suggestions on a projection/index which would allow me to do this?

Thanks

#2

It would be sufficient to be able to resolve an event based on a timestamp then read before or after that event depending on requirements.

I would imagine this would involve indexing the timestamp… would that make sense?

#3

Its easy to do externally … It will be easy to do internally very soon, what time frames are you looking at?

Greg

#4

Thanks Greg,

Excuse the noob question, still getting my head around indexing but its starting to make sense

If I wanted to execute this query, would I need to create an adhoc index where I would test the event timestamp against my from and to values, then if it is a match call linkTo to generate the result set?

I simply experimenting with EventStore for implementing this log so no immediate time frame.

#5

I guess i could create an index which would create a stream for each minute of the log. then I could read the relevant streams within my date range and trim off the entries which don’t apply for the seconds/milliseconds in the range.

This however would created a ton of streams… am I missing the point here?

#6

You can’t do this internally, yet. Event store is getting lucene indexing fairly soon. Or you can do thi externally in whatever read model you like.

#7

Ok I understand, thanks

#8

A stream per minute is actually no worries at all (streams are basically free we have a system with 100s of millions)

#9

Thanks, my only concern with this approach would be that I would need to truncate the messages at some point. This is easy enough to do using $maxCount however I will be left with tons of stream with links to missing events. I will wait for the Lucene style queries, sounds promising.