jayzhan211 commented on code in PR #10445:
URL: https://github.com/apache/datafusion/pull/10445#discussion_r1597326136


##########
datafusion/expr/src/logical_plan/builder.rs:
##########
@@ -188,37 +181,61 @@ impl LogicalPlanBuilder {
                     n_cols
                 );
             }
-            field_types = row
-                .iter()
-                .enumerate()
-                .map(|(j, expr)| {
-                    if let Expr::Literal(ScalarValue::Null) = expr {
-                        nulls.push((i, j));
-                        Ok(field_types[j].clone())
-                    } else {
-                        let data_type = expr.get_type(&empty_schema)?;
-                        if let Some(prev_data_type) = &field_types[j] {
-                            if prev_data_type != &data_type {
-                                return plan_err!("Inconsistent data type 
across values list at row {i} column {j}. Was {prev_data_type} but found 
{data_type}")
+        }
+
+        let empty_schema = DFSchema::empty();
+        let mut field_types: Vec<DataType> = Vec::with_capacity(n_cols);
+        for j in 0..n_cols {
+            let mut common_type: Option<DataType> = None;
+            for (i, row) in values.iter().enumerate() {
+                let value = &row[j];
+                if let Expr::Literal(ScalarValue::Null) = value {
+                    continue;
+                } else {
+                    let data_type = value.get_type(&empty_schema)?;
+                    if let Some(prev_type) = common_type {
+                        // get common type of each row values.
+                        match comparison_coercion(&data_type, &prev_type) {

Review Comment:
   The concern is that I'm not sure `comparison_coercion` is suitable for 
constructing Value, it is an easily misused `coercion`. I would recommend we 
start by adding numeric coercion for example the integer to floating. We can 
add more and more coercion rules covered with the test. If eventually, it is 
the same as `comparison_coercion`, then we can replace it.  But, I think we 
should start with simple.



-- 
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: github-unsubscr...@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For additional commands, e-mail: github-h...@datafusion.apache.org

Reply via email to