Question

SELECT [Cole_Crops], 
       PERCENT_RANK() OVER (PARTITION BY HUC02 
                            ORDER BY [Cole_Crops]) AS Cole_Crops_PCT_RANK], 
       [Row_Crops], 
       PERCENT_RANK() OVER (PARTITION BY HUC02 
                            ORDER BY [Row_Crops]) AS Row_Crops_PCT_RANK]            
FROM V012_CDLMAX_09

如何在具有多个要排名的列的查询中排除NULL值以影响PERCENT_RANK？

SQL Server 2012 PERCENT_RANK()将NULLS视为分发中的实际值。如果有一列要排名，您只需添加一个where子句（即where Cole_Crops IS NOT NULL）。

Answer 1

这是一个简单的修复，在你的分区子句中使用case语句。

CASE 
    WHEN COLUMN_1 IS NULL THEN NULL 
    ELSE PERCENT_RANK() OVER 
        (PARTITION BY 
            CASE 
                WHEN COLUMN_1 IS NULL THEN 0 
                ELSE 1 
            END 
         ORDER BY COLUMN_1) 
END AS RANK_COLUMN_1

这将分别对空值和非空值进行排名，但包装案例语句将空值显示为空值。

问题解决了！

Answer 2

您可以在Percent_Rank ORDER BY中使用case语句。您可以根据需要创建NULL值。下面是将所有NULL值视为0

SELECT [Cole_Crops], 
       PERCENT_RANK() OVER (PARTITION BY HUC02 
                            ORDER BY CASE WHEN [Cole_Crops] IS NULL THEN 0 ELSE [Cole_Crops] END) AS Cole_Crops_PCT_RANK], 
       [Row_Crops], 
       PERCENT_RANK() OVER (PARTITION BY HUC02 
                            ORDER BY CASE WHEN [Row_Crops] IS NULL THEN 0 ELSE [Row_Crops] END) AS Row_Crops_PCT_RANK]            
FROM V012_CDLMAX_09

Answer 3

您可以尝试在PARTITION BY子句中添加条件：

SELECT [Cole_Crops], 
   PERCENT_RANK() OVER (PARTITION BY [Cole_Crops] is not null, HUC02 
                        ORDER BY [Cole_Crops]) AS Cole_Crops_PCT_RANK], 
   [Row_Crops], 
   PERCENT_RANK() OVER (PARTITION BY [Row_Crops] is not null, HUC02 
                        ORDER BY [Row_Crops]) AS Row_Crops_PCT_RANK]
FROM V012_CDLMAX_09

Answer 4

对韦斯的回答稍作修改。应该是：

CASE 
    WHEN COLUMN_1 IS NULL THEN NULL 
    ELSE PERCENT_RANK() OVER 
        (PARTITION BY 
             HUC02
            END 
         ORDER BY COLUMN_1) 
END AS RANK_COLUMN_1

SQL Server 2012 PERCENT_RANK（）排除NULLS

4 个答案: