Github user juliuszsompolski commented on the issue:
https://github.com/apache/spark/pull/23152
While at it, could we kill one more potential for a bug?
In `FilterEstimation.evaluateBinaryForTwoColumns` there is a
```
attrLeft.dataType match {
case StringType | BinaryType =>
// TODO: It is difficult to support other binary comparisons for
String/Binary
// type without min/max and advanced statistics like histogram.
logDebug("[CBO] No range comparison statistics for String/Binary
type " + attrLeft)
return None
case _ =>
}
```
Could we change
```
case _ =>
if (!colStatsMap.hasMinMaxStats(attrLeft)) {
logDebug("[CBO] No min/max statistics " + attrLeft)
return None
}
if (!colStatsMap.hasMinMaxStats(attrRight)) {
logDebug("[CBO] No min/max statistics " + attrRight)
return None
}
```
This is one more place that later does
```
val statsIntervalLeft = ValueInterval(colStatLeft.min, colStatLeft.max,
attrLeft.dataType)
.asInstanceOf[NumericValueInterval]
...
val statsIntervalRight = ValueInterval(colStatRight.min,
colStatRight.max, attrRight.dataType)
.asInstanceOf[NumericValueInterval]
```
assuming that min/maxes are present, and could therefore also hint the
ClassCastException.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
