If the CH cluster is sharded and replicated, would the SQL query be executed on each cluster node? If so, wouldn't it be not good for an overall cluster performance?
is it the same as with incremental views that it depends on first table in join? and is only trigger for it? we needed to create a chain of mv's to get for whole view refreshed
The trigger for these is time based, so at the time it runs it will join whatever data is in each table. So not the same as for incremental MVs where the trigger is data being ingested into a source table
If the CH cluster is sharded and replicated, would the SQL query be executed on each cluster node? If so, wouldn't it be not good for an overall cluster performance?
is it the same as with incremental views that it depends on first table in join? and is only trigger for it? we needed to create a chain of mv's to get for whole view refreshed
The trigger for these is time based, so at the time it runs it will join whatever data is in each table.
So not the same as for incremental MVs where the trigger is data being ingested into a source table
Super interesting features, but it’s impossible to follow this video, had to watch it at 0.5x
Is it ok to put a really heavy SQL in Refreshable MV that runs for example 10 minutes and uses a lot of system resources?