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
To filter a list of events down based on person properties, we join on the person table. However, this means that when we fan out the queries to other nodes, instead of them just responding with a count, they have to send over every single event, which we then join on the main node. This means transferring 10s of GBs of data.
We should move join into a subquery instead, so the nodes can respond with the total count instead.
The problem
To filter a list of events down based on person properties, we join on the person table. However, this means that when we fan out the queries to other nodes, instead of them just responding with a count, they have to send over every single event, which we then join on the main node. This means transferring 10s of GBs of data.
We should move join into a subquery instead, so the nodes can respond with the total count instead.
See this query for an overview
Example query before moving person to subquery
Example query after moving person to subquery
The text was updated successfully, but these errors were encountered: