Oracle create view error

[cameronclaero]

Bug description

Hi having an issue using the Allow create view as option in superset using oracle, I'm setting a specific schema, and when going to save the view, getting the following error message: Database error
oracle error: ORA-00903: invalid table name

Screenshots/recordings

No response

Superset version

5.0.0

Python version

3.10

Node version

16

Browser

Chrome

Additional context

No response

Checklist

• I have searched Superset docs and Slack and didn't find a solution to my problem.
• I have searched the GitHub issue tracker and didn't find a similar bug report.
• I have checked Superset's logs for errors and if I found a relevant Python stacktrace, I included it here as text in the "additional context" section.

[dosubot]

Hi @cameronclaero! I'm Dosu and I’m helping the superset team.

This Oracle error in Superset 5.0.0 is often caused by how Superset constructs the CREATE VIEW statement and handles schema/table naming for Oracle, which has strict requirements. Common causes include invalid or improperly quoted table/view names, reserved words, or exceeding Oracle's identifier length limits (Superset still uses a 30-character limit, which is outdated for newer Oracle versions) source.

To troubleshoot, enable debug logging in Superset and check the exact SQL being executed. Look for issues like missing schema qualification, reserved words not quoted with double quotes, or table/view names that violate Oracle's rules. If possible, try manually creating the view in Oracle using the same SQL to isolate the problem.

A known workaround is to recreate the dataset as a physical dataset instead of a virtual one—physical datasets use SQLAlchemy's inspector to fetch column types directly from the schema, which is more reliable for Oracle source.

If the error persists, review your schema and table/view names for Oracle compatibility, and consider upgrading Superset if you need support for longer identifiers or improved schema handling. There are no recent fixes for this in Superset 5.0.0, so manual adjustments or workarounds are currently the best option.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[cameronclaero]

@dosu where is the code that does that parsing, I will try and submit a pr to fix it.

[dosubot]

The main code for parsing and constructing SQL statements—including CREATE VIEW for Oracle—is in superset/sql/parse.py. The SQLStatement class (and related SQLScript class) uses sqlglot to parse SQL, extract tables, and rewrite statements as CREATE TABLE or CREATE VIEW AS, with dialect-specific logic for Oracle and others. Schema and table naming are handled by the Table dataclass in the same file, which ensures proper qualification and quoting. If you want to adjust how Superset builds or parses CREATE VIEW statements for Oracle (e.g., to fix quoting or schema handling), start with superset/sql/parse.py, especially the SQLStatement and Table classes. You may also want to review OracleEngineSpec for engine-specific quoting rules and identifier limits.
Source: superset/sql/parse.py

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[cameronclaero]

@dosu is there a way to log what sql statement is attempted to run on the create view? using kubernetes helm chart to deploy. eg if sqlalchemy is being used some type of logging being turned on? i already have the env var DEBUG on but that does not output much

[dosubot]

Yes, you can log the exact SQL statements (including CREATE VIEW) that Superset sends to your database, but setting the DEBUG environment variable alone is not enough. You need to enable SQLAlchemy's SQL logging.

For a Helm/Kubernetes deployment, the recommended approach is to use the configOverrides field in your values.yaml to append to superset_config.py. Add the following override to enable SQL statement logging:

configOverrides:
  enable_sqlalchemy_echo: |
    SQLALCHEMY_ENGINE_OPTIONS = {"echo": True}

Alternatively, you can set this in a custom superset_config.py if you mount one.

If you want to increase overall log verbosity, you can also set:

extraEnv:
  SUPERSET_LOG_LEVEL: "debug"

With these settings, all SQL statements (including CREATE VIEW) will be logged to stdout/stderr, which means you can see them in your pod logs via kubectl logs or your cluster's log aggregation. If you want persistent logs, you can also enable file logging in superset_config.py, but for most Kubernetes setups, stdout/stderr is preferred reference reference.

Remember to remove or revert these changes after debugging, as SQL logging can be very verbose and may impact performance.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}