rdettai opened a new issue #1027: URL: https://github.com/apache/arrow-datafusion/issues/1027
**Describe the bug** The [`BallistaContext.register_table(name,dataframe)`](https://github.com/apache/arrow-datafusion/blob/db305dadea458a26723fa53d85463155247d39f9/ballista/rust/client/src/context.rs#L200) method can register any type of dataframe. But when serializing the plan to send it to the scheduler, it will only work if value wrapped [if the `DfTableAdapter` is a `LogicalPlan::TableScan`](https://github.com/apache/arrow-datafusion/blob/db305dadea458a26723fa53d85463155247d39f9/ballista/rust/core/src/serde/logical_plan/to_proto.rs#L733-L744). **To Reproduce** Steps to reproduce the behavior: - create a dataframe with any of the `BallistaContext.read_xxx()` method - add another operation like a projection - register the result with `BallistaContext.register_table()` - build an sql query that uses the registered table - running this query will raise an error **Expected behavior** Two possibilities: - either we can only register instances of `TableProvider` instead of `DataFrame` (just as in Datafusion) - we add an extra step that unwraps `DfTableAdapter` before serialization -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
