Fixed extremes calculation in presense of NaNs [#METR-22882].

dbms/include/DB/Columns/ColumnVector.h

@@ -120,6 +120,27 @@ template <> inline UInt64 unionCastToUInt64(Float32 x)
 }
 
 
+/// To be sure, that this function is zero-cost for non-floating point types.
+template <typename T>
+inline bool isNaN(T x)
+{
+	return std::is_floating_point<T>::value ? isnan(x) : false;
+}
+
+
+template <typename T>
+typename std::enable_if<std::is_floating_point<T>::value, T>::type NaNOrZero()
+{
+	return std::numeric_limits<T>::quiet_NaN();
+}
+
+template <typename T>
+typename std::enable_if<!std::is_floating_point<T>::value, T>::type NaNOrZero()
+{
+	return 0;
+}
+
+
 /** Шаблон столбцов, которые используют для хранения простой массив.
   */
 template <typename T>
@@ -435,16 +456,35 @@ public:
 			return;
 		}
 
-		T cur_min = data[0];
+		bool has_value = false;
-		T cur_max = data[0];
 
-		for (size_t i = 1; i < size; ++i)
+		/** Skip all NaNs in extremes calculation.
+		  * If all values are NaNs, then return NaN.
+		  * NOTE: There exist many different NaNs.
+		  * Different NaN could be returned: not bit-exact value as one of NaNs from column.
+		  */
+
+		T cur_min = NaNOrZero<T>();
+		T cur_max = NaNOrZero<T>();
+
+		for (const T x : data)
 		{
-			if (data[i] < cur_min)
+			if (isNaN(x))
-				cur_min = data[i];
+				continue;
 
-			if (data[i] > cur_max)
+			if (!has_value)
-				cur_max = data[i];
+			{
+				cur_min = x;
+				cur_max = x;
+				has_value = true;
+				continue;
+			}
+
+			if (x < cur_min)
+				cur_min = x;
+
+			if (x > cur_max)
+				cur_max = x;
 		}
 
 		min = typename NearestFieldType<T>::Type(cur_min);

dbms/tests/queries/0_stateless/00402_nan_and_extremes.reference

@@ -0,0 +1,48 @@
+3
+1
+2
+
+1
+3
+nan
+1
+2
+
+1
+2
+3
+nan
+2
+
+2
+3
+3
+1
+nan
+
+1
+3
+nan
+nan
+2
+
+2
+2
+nan
+1
+nan
+
+1
+1
+3
+nan
+nan
+
+3
+3
+nan
+nan
+nan
+
+nan
+nan

dbms/tests/queries/0_stateless/00402_nan_and_extremes.sql

@@ -0,0 +1,8 @@
+SELECT arrayJoin([3, 1, 2]) SETTINGS extremes = 1;
+SELECT arrayJoin([nan, 1, 2]) SETTINGS extremes = 1;
+SELECT arrayJoin([3, nan, 2]) SETTINGS extremes = 1;
+SELECT arrayJoin([3, 1, nan]) SETTINGS extremes = 1;
+SELECT arrayJoin([nan, nan, 2]) SETTINGS extremes = 1;
+SELECT arrayJoin([nan, 1, nan]) SETTINGS extremes = 1;
+SELECT arrayJoin([3, nan, nan]) SETTINGS extremes = 1;
+SELECT arrayJoin([nan, nan, nan]) SETTINGS extremes = 1;