在SQL中联接表，其中新列名基于另一列的联接表值

Question

I have 2 tables and I'm trying to join contents of the second table, into the first. In essence i want to end up with a list of items that contains all fields from the first table, and all relevant rows in the second table, names after column content in the second table.

Table A

+----+--------+---------------------+--------+
| ID | Title  |        Date         | Etc... |
+----+--------+---------------------+--------+
|  1 | QWERTY | 0000-00-00:00:00:00 |    ... |
|  2 | TEST   | 0000-00-00:00:00:00 |    ... |
|  3 | ASDF   | 0000-00-00:00:00:00 |    ... |
|  4 | IOP    | 0000-00-00:00:00:00 |    ... |
+----+--------+---------------------+--------+

Table B

+----+------------+------+------------+
| ID | Table_A_ID | Key  |   Value    |
+----+------------+------+------------+
|  1 |          1 | Name | John       |
|  2 |          2 | Name | Brad       |
|  3 |          1 | DOB  | 2000-01-16 |
+----+------------+------+------------+

So I'd like to be able to Select * from Table A and somehow end up with the following

+----+--------+---------------------+--------+------+------------+
| ID | Title  |        Date         | Etc... | Name |    DOB     |
+----+--------+---------------------+--------+------+------------+
|  1 | QWERTY | 0000-00-00:00:00:00 |    ... | John | 2000-01-16 |
|  2 | TEST   | 0000-00-00:00:00:00 |    ... | Brad |            |
|  3 | ASDF   | 0000-00-00:00:00:00 |    ... |      |            |
|  4 | IOP    | 0000-00-00:00:00:00 |    ... |      |            |
+----+--------+---------------------+--------+------+------------+

The key thing is that i don't know what the keys are going to be

Unfortunately I have no idea what this is called to google it, I've been looking through the inner, left, right joins and these don't seem to do what I'm trying to do.

Answer 1

This can be achieved with two JOINs on TableB, like :

SELECT
    ta.id,
    ta.date,
    ta.etc,
    tb.value Name,
    td.value DOB
FROM
    tableA ta
    INNER JOIN tableB tb ON tb.table_a_id = tb.id AND tb.key = 'Name'
    INNER JOIN tableB td ON td.table_a_id = ta.id AND td.key = 'DOB'

With this query, only records that do have a Name and a DOB in TableB will be displayed. If you want to avoid filtering out these "incomplete" records, you can use LEFT JOIN instead of INNER JOIN.

Answer 2

Select a.*, 
(select TOP 1  b.DOB from b where b.table_a_id = a.id and b.key = 'dob'), 
(select TOP 1 b.ame from b where b.table_a_id = a.id and b.key = 'name') 
from a

You have to use sub-selects to achieve that results

Answer 3

You're looking to pivot the keyed values. You can do that manually like this:

select a.*
     , max(case b.key when 'Name' then b.value end) Name
     , max(case b.key when 'DOB' then b.value end) DOB
  from TableA a
  left join TableB b
    on a.id = b.table_a_id
  group by a.id
      , a.title
      , a.date
      , a.etc

The above query will work if you only have one of each attribute for a given record in TableA, otherwise it will give you the largest value. If you want multiple values, you can generate a sequence for each attribute as shown in the subselect named B below:

select a.ID
     , b.Seq
     , a.Title
     , a.Date
     , a.Etc
     , max(case b.key when 'Name' then b.value end) Name
     , max(case b.key when 'DOB' then b.value end) DOB
  from TableA a 
  left join (select @Seq:=case
                       when @aid=b.Table_A_ID and @key=b.key
                       then @Seq + 1
                       else 1
                     end Seq
                   , @aid:=b.Table_A_ID Table_A_ID
                   , @key:=b.key `Key`
                   , b.value
                from tableB b
               order by table_a_id, b.key) b
     on a.id = b.table_a_id
  group by a.id, a.title, a.date, a.etc, b.Seq;

This is different than @GMB's response in that it will return each attribute only once with nulls in rows where there are fewer values per attribute whereas @GMB's solution will return the CROSS product of all attributes for a given row.