phillipleblanc commented on code in PR #12609:
URL: https://github.com/apache/datafusion/pull/12609#discussion_r1774284699
##########
datafusion/sql/src/unparser/expr.rs:
##########
@@ -952,13 +952,25 @@ impl Unparser<'_> {
Ok(ast::Expr::Value(ast::Value::Number(f.to_string(), false)))
}
ScalarValue::Float16(None) =>
Ok(ast::Expr::Value(ast::Value::Null)),
- ScalarValue::Float32(Some(f)) => {
- Ok(ast::Expr::Value(ast::Value::Number(f.to_string(), false)))
- }
+ ScalarValue::Float32(Some(f)) => Ok(ast::Expr::Cast {
+ kind: ast::CastKind::Cast,
+ expr: Box::new(ast::Expr::Value(ast::Value::Number(
+ f.to_string(),
+ false,
+ ))),
+ data_type: self.arrow_dtype_to_ast_dtype(&DataType::Float32)?,
+ format: None,
+ }),
Review Comment:
This solves the problem of keeping the float literal types intact when
generating SQL to execute against remote systems, but it also has the side
effect of making the generated SQL a bit less human readable. Instead of just
calling `.to_string()` on the float literal, which chops off the decimal
portion if it doesn't exist - could we format it to always contain a decimal
portion? i.e. `3` -> `3.0`. That should solve the problem and keep the
generated SQL easier to understand by humans.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
