cast expression may cause duplicate column name error

[gaoxinge]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.

This problem origin from dask-contrib/dask-sql#948, which can be minimally reproduced by

SELECT 10, cast(10 as tinyint)

The sql above will cause the duplicate column name error. And this is because

• cast use the expr as column name

https://github.com/apache/arrow-datafusion/blob/224c682101949da57aebc36e92e5a881ef3040d4/datafusion/expr/src/expr.rs#L1203-L1210

• same column name is not valid

https://github.com/apache/arrow-datafusion/blob/224c682101949da57aebc36e92e5a881ef3040d4/datafusion/expr/src/logical_plan/builder.rs#L925-L948

Describe the solution you'd like

There are two soluntions here:

1. rename the column name of cast expression

For example, mysql use the whole cast expression as column name, like CAST(Int64(10) AS Int8) instead of Int64(10). In this way, the column name of cast expression will be more precise, and also this can bypass the duplicate column name error.

2. remove the unique column name validation

In mysql and sql server, query support same column name. So I think we can remove the unique column name validation to make same column name coexist.

[alamb]

I think @andygrove and @xudong963 may have had thoughts on this topic previously

[alamb]

Also, possibly related -- #4723 from @comphead

[comphead]

Thanks @gaoxinge for reporting this, I'm planning to start work on this as @andygrove agreed here #3990 (comment)

[gaoxinge]

@alamb @comphead Thx for reply, and I see the progress and direction of this issue.