Question

我在vertica db中有以下表格：

+-----+-------+-------+-------+-------+
| Tid | Item1 | Item2 | Item3 | Item4 |
+-----+-------+-------+-------+-------+
|   1 | A     | B     | C     | D     |
|   2 | A     | E     | B     | D     |
|   3 | D     | A     | 5     | C     |
|   4 | B     | 1     | E     | A     |
+-----+-------+-------+-------+-------+

我的目标是以字母数字排序行，所有条目都是字符串。结果看起来像这样：

+-----+-------+-------+-------+-------+
| Tid | Item1 | Item2 | Item3 | Item4 |
+-----+-------+-------+-------+-------+
|   1 | A     | B     | C     | D     |
|   2 | A     | B     | D     | E     |
|   3 | 5     | A     | C     | D     |
|   4 | 1     | A     | B     | E     |
+-----+-------+-------+-------+-------+

用法：我使用sql生成apriori算法，为了生成新的候选人，我想使用F_k-1加入，在我的理解中需要按照上面的要求进行排序。

感谢您的帮助。

Answer 1

我无法访问Vertica，但这应该可以胜任（虽然我不能承认这是Vertica最有效的方式/最干净的代码）

select      Tid
           ,min (case rn when 1 then item end)  as Item1
           ,min (case rn when 2 then item end)  as Item2
           ,min (case rn when 3 then item end)  as Item3
           ,min (case rn when 4 then item end)  as Item4

from       (select      Tid,item
                       ,row_number () over (partition by Tid order by item) as rn

            from       (            select Tid ,Item1 as item from t
                        union all   select Tid ,Item2         from t
                        union all   select Tid ,Item3         from t
                        union all   select Tid ,Item4         from t
                        ) t
            ) t

group by    Tid

+-----+-------+-------+-------+-------+
| tid | item1 | item2 | item3 | item4 |
+-----+-------+-------+-------+-------+
| 1   | A     | B     | C     | D     |
+-----+-------+-------+-------+-------+
| 2   | A     | B     | D     | E     |
+-----+-------+-------+-------+-------+
| 3   | 5     | A     | C     | D     |
+-----+-------+-------+-------+-------+
| 4   | 1     | A     | B     | E     |
+-----+-------+-------+-------+-------+

Sql：按字母顺序排序行条目

1 个答案: