当列数据位于带有鉴别器列的单列中时，SQL Pivot

Question

我有一组数据代表人们如何对各种特征进行排名。我试图将数据按两列进行透视，但是，列数据包含在单个列下，单独的列充当鉴别器。

这里是场景。人们得到一份包含三个特征的列表，并要求他们按照对他们的重要性排序。他们获得了两个列表，每个列表包含三个特征，并要求他们将每个列表从最重要到最不重要排列为 1 到 3。

+----------+-------+----------------+---------------+------+
| RecordNo |  Name | QuestionNumber | QuestionGroup | Rank |
+----------+-------+----------------+---------------+------+
|     1    |  Bob  |        1       |       1       |   2  |
|     2    |  Bob  |        2       |       1       |   1  |
|     3    |  Bob  |        3       |       1       |   3  |
|     4    |  Bob  |        1       |       2       |   1  |
|     5    |  Bob  |        2       |       2       |   2  |
|     6    |  Bob  |        3       |       2       |   3  |
|     7    | Sally |        1       |       1       |   3  |
|     8    | Sally |        2       |       1       |   2  |
|     9    | Sally |        3       |       1       |   1  |
|    10    | Sally |        1       |       2       |   1  |
|    11    | Sally |        2       |       2       |   3  |
|    12    | Sally |        3       |       2       |   2  |
+----------+-------+----------------+---------------+------+

我想最终得到的是数据的 PIVOT，所以它看起来像这样..

+----------+-------+-----------+-----------+-----------+------------+------------+------------+
| RecordNo |  Name | Question1 | Question2 | Question3 | Question 1 | Question 2 | Question 3 |
|          |       |  Group 1  |  Group 1  |  Group 1  |   Group 2  |   Group 2  |   Group 2  |
+----------+-------+-----------+-----------+-----------+------------+------------+------------+
|     1    |  Bob  |     2     |     1     |     3     |      1     |      2     |      3     |
|     2    | Sally |     3     |     2     |     1     |      1     |      3     |      2     |
+----------+-------+-----------+-----------+-----------+------------+------------+------------+

我知道如何在多个列 great article on it here 上进行 Pivot，但我无法弄清楚当数据位于由鉴别器列 (QuestionGroup) 分隔的同一列 (QuestionNumber) 中时如何进行 Pivot。

我还创建了一个在线表here

Answer 1

我认为最简单的透视方法是条件聚合：

select name,
       max(case when questiongroup = 1 and questionnumber = 1 then rank end) as q_1_1,
       max(case when questiongroup = 1 and questionnumber = 2 then rank end) as q_1_2,
       max(case when questiongroup = 1 and questionnumber = 3 then rank end) as q_1_3,
       max(case when questiongroup = 2 and questionnumber = 1 then rank end) as q_2_1,
       max(case when questiongroup = 2 and questionnumber = 2 then rank end) as q_2_2,
       max(case when questiongroup = 2 and questionnumber = 3 then rank end) as q_2_3
from t
group by name;