Apache Doris
Apache Doris is its own database type in QueryDesk. It speaks the MySQL protocol, so the schema sync, data protection and peer reviews work as they do on MySQL, with Doris's catalogs and SQL. QueryDesk is tested against Doris 2.1.
Connect a cluster
Add the database as in Connect a database, with these values:
| Field | Value |
|---|---|
| Database type | Apache Doris |
| Database IP or hostname | A frontend (FE) node, or the load balancer in front of them. |
| Port | 9030, the FE's MySQL port. |
| Database name | The database to query, or catalog.database for an external catalog, for example hive.sales. A bare name is in the internal catalog. |
A mysql:// or doris:// connection string pasted under Update with connection string keeps the Apache Doris type.
Click Test Connection under the credential to check the connection.
What works differently
- One catalog per database. The schema sync reads the catalog the database name gives, and data protection refuses a statement that switches catalog (
SWITCH,SET CATALOG) or reads another catalog's tables. - Primary keys are a Unique Key table's key columns. Doris keeps no foreign keys or secondary indexes, so none are synced.
- AI read-only rests on QueryDesk's parser alone: Doris accepts
SET TRANSACTION READ ONLYbut writes anyway. The parser refuses thes3(),hdfs(),local()andhttp()table functions, which read files and URLs, andquery(), which runs SQL on a catalog's source. Give the AI's credentials onlySELECT_PRIV. - Workflow variables are written into the statement as escaped string literals, since Doris binds no parameters.
- SHOW is refused under a policy, since
SHOW PROCcan print other sessions' SQL.