I am quite new to ElasticSearch, most of what I need to seems to be easily
supported however I have hit one problem - specifically searching by
What do I mean by this? I'll explain a simplified scenario.
I have indexed events each with an associated date-time. I need to
discover those events that occur within a given time period of each other.
Ideally this time period would be arbitrary and specified at search time.
The "of each other" part is key here. I cannot just, say:
- Aggregate by week or month, as an event that occurs at the end of one
month is within a month of one which occurs at the start of the next month
but would fall into a different bucket.
- Do a date range based search, as I do not have a fixed range to search.
To describe what I need in a different way, in SQL I'd do something like:
SELECT eventid, eventdate from Events t1 where EXISTS(
SELECT 1 FROM Events t2 WHERE
t1.eventid <> t2.eventid AND
dateadd( day, -30, t2.eventdate )
dateadd( day, 30, t2.eventdate )
The actual scenario is more complex - each event has a type and what I need
ultimately is to be able to answer questions of akin to:
"Find events of type X that occur with in 2 days of an event of type Y"
"Find events of type X that occur with in 2 days of an event of type Y and
of type Z"
Each event will be nested in/a child of, a parent record and I am only
interested in the temporal proximity of events with the same parent. The database
has a total of 10^9 events, each parent may have the order of 10^3
associated events. The use case is search heavy with ingests of deltas
I can munge the data on import in any way that would help. I have had a
couple of ideas on how to tackle the problem but neither are satisfactory.
I wondered whether there is a standard way to tackle this kind of
requirement in ElasticSearch and whether anyone else had run up against it.
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to email@example.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/31bc7955-2043-49dd-865d-1dbc048a6dde%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.