Fokko commented on code in PR #36846:
URL: https://github.com/apache/arrow/pull/36846#discussion_r1307101413


##########
cpp/src/arrow/type.h:
##########
@@ -397,14 +397,67 @@ class ARROW_EXPORT Field : public detail::Fingerprintable,
   /// \brief Options that control the behavior of `MergeWith`.
   /// Options are to be added to allow type conversions, including integer
   /// widening, promotion from integer to float, or conversion to or from 
boolean.
-  struct MergeOptions {
+  struct ARROW_EXPORT MergeOptions : public 
util::ToStringOstreamable<MergeOptions> {
     /// If true, a Field of NullType can be unified with a Field of another 
type.
     /// The unified field will be of the other type and become nullable.
     /// Nullability will be promoted to the looser option (nullable if one is 
not
     /// nullable).
     bool promote_nullability = true;
 
+    /// Allow a decimal to be unified with another decimal of the same
+    /// width, adjusting scale and precision as appropriate. May fail
+    /// if the adjustment is not possible.
+    bool promote_decimal = false;
+
+    /// Allow a decimal to be promoted to a float. The float type will
+    /// not itself be promoted (e.g. Decimal128 + Float32 = Float32).
+    bool promote_decimal_to_float = false;
+
+    /// Allow an integer to be promoted to a decimal.
+    ///
+    /// May fail if the decimal has insufficient precision to
+    /// accommodate the integer. (See increase_decimal_precision.)
+    bool promote_integer_to_decimal = false;
+
+    /// Allow an integer of a given bit width to be promoted to a
+    /// float; the result will be a float of an equal or greater bit
+    /// width to both of the inputs.
+    bool promote_integer_to_float = false;

Review Comment:
   Updated the docstring to:
   ```
   /// Allow an integer of a given bit width to be promoted to a
   /// float; the result will be a float of an equal or greater bit
   /// width to both of the inputs. Examples:
   ///  - int8 + float32 = float32
   ///  - int32 + float32 = float64
   ///  - int32 + float64 = float64
   /// Because an int32 cannot always be represented exactly in the 
   /// 24 bits of a float32 mantissa.
   ```
   Let me know if you had an other place in mind to document this.



-- 
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]

Reply via email to