Tim Fox
Tim Fox
We can consider sources which get their data from an external database table (e.g. in MySQL or in TiDB) instead of from a Kafka topic. A CDC tool such as...
We should allow quotas and limits to be specified on a per user level. Quotas could include such things as: * Max number of queries that can be executed per...
Currently we use push queries internally in order to compute rows to update a materialized view. We could also allow users to execute push queries from their applications and have...
We should consider implementing remote table or aggregate functions. These functions can be used in the SQL that defines a MV and results in a remote call via gRPC to...
Some operations - such as filling an MV or an index are executed on nodes which are leaders for the shards. This ensures a) The process is only executed on...
Currently the number of nodes and shards in a cluster is fixed and cannot change. We can keep the number of shards in a cluster fixed but we should allow...
A push query can be suffixed with emit after delay , e.g. emit after delay 5s. In which case updates are not emitted every time the results change, but they...
If a delete arrives at an push aggregator then the value needs to be reversed out of the current aggregation. In the case of some aggregations such as sum or...
The LATEST aggregate function should take the latest value of the column in the aggregation. The EARLIEST should take the first value. An implementation of EARLIEST already exists named as...