You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
The basic problem is: I have a stream of events, ie user behavior on a website, I want to find out how many of the streams have a user visiting a page, then scrolling within a given period of time, then doing another thing, without a page switch in the meantime...
I might then want to count how often this happened, or see user trends or whatever else.
This sort of event stream analysis is very difficult in SQL, there are efforts to work this into new syntax for SQL namely the match_recognize keyword: https://modern-sql.com/feature/match_recognize, but that only exists in certain Oracle implementations for now and I find the syntax somewhat confusing, though very powerful.
this is also very powerful for things like log analysis and the like.
Describe the solution you'd like
Extend the timeseries API to work with event data and provide a series of functions for analyzing / fingerprinting event series (perhaps it's actually a separate type called an eventseries?). You could then provide a regex like pattern matching language that has a part of it describing the length of time that can be between events etc.
Additional context
This comes out of a conversation with a data analyst friend who described some difficult problems they were facing.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
The basic problem is: I have a stream of events, ie user behavior on a website, I want to find out how many of the streams have a user visiting a page, then scrolling within a given period of time, then doing another thing, without a page switch in the meantime...
I might then want to count how often this happened, or see user trends or whatever else.
This sort of event stream analysis is very difficult in SQL, there are efforts to work this into new syntax for SQL namely the match_recognize keyword: https://modern-sql.com/feature/match_recognize, but that only exists in certain Oracle implementations for now and I find the syntax somewhat confusing, though very powerful.
this is also very powerful for things like log analysis and the like.
Describe the solution you'd like
Extend the timeseries API to work with event data and provide a series of functions for analyzing / fingerprinting event series (perhaps it's actually a separate type called an eventseries?). You could then provide a regex like pattern matching language that has a part of it describing the length of time that can be between events etc.
Additional context
This comes out of a conversation with a data analyst friend who described some difficult problems they were facing.
The text was updated successfully, but these errors were encountered: