使用 T-SQL 转置两个文本列

Question

我可以使用 T-SQL 查询来转置文本数据行吗：

像这样的专栏：

每个人可以参加的培训次数各不相同？

Answer 1

希望这对您有所帮助。任何ID超过3个培训课程，多余的培训课程将不会出现。

SELECT OrgDefinedID
    , max(CASE WHEN row_num = 1 THEN attended END) Attendance1
    , max(CASE WHEN row_num = 1 THEN training END) Training1
    , max(CASE WHEN row_num = 2 THEN attended END) Attendance2
    , max(CASE WHEN row_num = 2 THEN training END) Training2
    , max(CASE WHEN row_num = 3 THEN attended END) Attendance3
    , max(CASE WHEN row_num = 3 THEN training END) Training3

FROM
(SELECT row_number() over (partition by OrgDefinedID order by Attended, Training) as row_num, *  
FROM T1) x 
group by OrgDefinedID

max() 在做什么？我的第一次尝试是这样

SELECT OrgDefinedID
    , (CASE WHEN row_num = 1 THEN attended END) Attendance1
    , (CASE WHEN row_num = 1 THEN training END) Training1
    , (CASE WHEN row_num = 2 THEN attended END) Attendance2
    , (CASE WHEN row_num = 2 THEN training END) Training2
    , (CASE WHEN row_num = 3 THEN attended END) Attendance3
    , (CASE WHEN row_num = 3 THEN training END) Training3

FROM
(SELECT row_number() over (partition by OrgDefinedID order by Attended, Training) as row_num, *  
FROM T1) x

快到了。我们只需要将多行折叠成一行

OrgDefinedID	Attendance1	Training1	Attendance2	Training2	Attendance3	Training3
13076	Yes	CPIE Online Training Studio	null	null	null	null
13076	null	null	Yes	AIDI	null	null
12505478	Yes	Best Practices 3, 9AM	null	null	null	null
12505478	null	null	Yes	Work Smarter II, #4	null	null
12505478	null	null	null	null	Yes	Rubric Design 2

您可以看到，在给定的 OrgDefinedID 中，每一列都有一个非空值，而该列中的其他值为空。因此，如果我们按 OrgDefinedID 分组并应用 max()，它将选择我们想要的非空值。请注意，像 max() 这样的聚合函数会忽略空值。所以添加 max() 函数并按 OrgDefinedID 分组，我们得到：

OrgDefinedID	Attendance1	Training1	Attendance2	Training2	Attendance3	Training3
13076	Yes	CPIE Online Training Studio	Yes	AIDI	null	null
12505478	Yes	Best Practices 3, 9AM	Yes	Work Smarter II, #4	Yes	Rubric Design 2

使用 T-SQL 转置两个文本列

Use T-SQL to transpose two text columns

tsql

transpose

row