How to specify the constant type（int32， int64...） in an SQL expression； as if the default is int64， how to specify int32

[hrh007]

when in write sql select column1 = 123, and column1 type is int32, it doesn't work;

log is : Error: Internal("The type of Int32 = Int64 of binary physical should be same")

[alamb]

Hi @hrh007 -- that definitely sounds like a bug - if you have a reproducer I would love to see it

A workaround might be to put an explicit cast in:

For example this would cast 123 to bigint (Int64):

select column1 = 123::bigint

The SQL --> Arrow data type mapping for DataFusion is on https://arrow.apache.org/datafusion/user-guide/sql/data_types.html

[hrh007]

And if use typecast, SQL is : select udf_xxx( column1 = 3015::int, ...)

Error: NotImplemented("Physical plan does not support logical expression; just used by udf udf_xxx( table5.column1 = Int32(3015), ) AS udtf_xxx(table5.column1 = Int64(3015))")

Physical plan and logical expression type mismatch

[alamb]

it sounds like a bug to me -- perhaps you can provide a reproducer so we can investigate it more carefully?

[hrh007]

Can my examples above be reproduced in your environment ？

[alamb]

I tried to reproduce it using datafusion-cli and I could not:

DataFusion CLI v26.0.0
❯ create table t as select 1::int as column1;
0 rows in set. Query took 0.031 seconds.
❯ select * from t;
+---------+
| column1 |
+---------+
| 1       |
+---------+
1 row in set. Query took 0.005 seconds.
❯ select arrow_typeof(column1) from t;
+-------------------------+
| arrow_typeof(t.column1) |
+-------------------------+
| Int32                   |
+-------------------------+
❯ select column1 = 123 from t;
+------------------------+
| t.column1 = Int64(123) |
+------------------------+
| false                  |
+------------------------+
1 row in set. Query took 0.005 seconds.

[hrh007]

My version is datafusion 14.0
Your example above is also correct in my cli env；
but when I write an rust example in datafusion source code : register_parquet, run a sql; the wrong will occur