[
https://issues.apache.org/jira/browse/FLINK-38042?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Etienne Chauchot updated FLINK-38042:
-------------------------------------
Description:
The Cassandra source connector allows the user to specify a query. This query
can contain filters on the provided table. So these filters are pushed down but
in some cases, they could lead to inefficient table scans:
- Filtering on non-indexed, non-key columns (primary key or clustering columns)
- not restricting on partition key leads the cluster to read all the partitions
We should detect such cases and warn the user.
was:
The Cassandra source connector allows the user to specify a query. This query
can contain filters on the provided table. So these filters are pushed down but
in some cases, they could lead to inefficient table scans:
- Filtering on non-indexed, non-key columns
- not restricting on partition key leads the cluster to read all the partitions
We should detect such cases and warn the user.
> Warn the user on full table scans
> ---------------------------------
>
> Key: FLINK-38042
> URL: https://issues.apache.org/jira/browse/FLINK-38042
> Project: Flink
> Issue Type: Improvement
> Components: Connectors / Cassandra
> Reporter: Etienne Chauchot
> Assignee: Etienne Chauchot
> Priority: Major
>
> The Cassandra source connector allows the user to specify a query. This query
> can contain filters on the provided table. So these filters are pushed down
> but in some cases, they could lead to inefficient table scans:
> - Filtering on non-indexed, non-key columns (primary key or clustering
> columns)
> - not restricting on partition key leads the cluster to read all the
> partitions
> We should detect such cases and warn the user.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
