HaoYang670 commented on code in PR #3669:
URL: https://github.com/apache/arrow-datafusion/pull/3669#discussion_r985236424
##########
datafusion/optimizer/src/simplify_expressions.rs:
##########
@@ -790,6 +791,37 @@ impl<'a, S: SimplifyInfo> ExprRewriter for Simplifier<'a,
S> {
right,
} if !info.nullable(&left)? && left == right => lit(1),
+ //
+ // Rules for Modulo
+ //
+
+ // A % null --> null
+ BinaryExpr {
+ left: _,
+ op: Modulo,
+ right,
+ } if is_null(&right) => *right,
+ // null % A --> null
+ BinaryExpr {
+ left,
+ op: Modulo,
+ right: _,
+ } if is_null(&left) => *left,
+ // A % 1 --> 0
+ BinaryExpr {
+ left,
+ op: Modulo,
+ right,
+ } if !info.nullable(&left)? && is_one(&right) => lit(0),
+ // A % 0 --> DivideByZero Error
+ BinaryExpr {
+ left,
+ op: Modulo,
+ right,
+ } if !info.nullable(&left)? && is_zero(&right) => {
+ return
Err(DataFusionError::ArrowError(ArrowError::DivideByZero))
Review Comment:
Hi @alamb @askoa, thank you for bringing up the discussion about the error
checking. Here is my opinion.
We should do as much error checking as possible at compile time, whatever
the error will be evaluated or not. Because this is much cheaper than checking
at runtime. We don't want a long query to panic at an error that could be
checked at compile time, such as `A % literal(0)`.
In Postgresql, `A % literal(0)` is checked at compiled time:
https://onecompiler.com/postgresql/3yhrjpgrr
--
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]
