将多行合并为一行MySQL

Question

Say I have two tables in a MySQL Database.

Table 1:

ID    Name
1     Jim
2     Bob
3     John

Table 2:

ID    key           value
1     address       "X Street"
1     city          "NY"
1     region        "NY"
1     country       "USA"
1     postal_code   ""
1     phone         "123456789"

When selecting rows from the database, is there any way to join rows from the second table as columns to the first table?

The desired result right from the MySQL query is:

ID    Name    address    city    region   country   postal_code   phone
1     Jim     X Street   NY      NY       USA       NULL          123456789
2     Bob     NULL       NULL    NULL     NULL      NULL          NULL
3     John    NULL       NULL    NULL     NULL      NULL          NULL

Thanks for any help!

Answer 1

This type of data transformation is known as a PIVOT. MySQL doesn't have a pivot function but you can replicate it using an aggregate function with a CASE expression:

select t1.id,
  t1.name,
  max(case when t2.`key` = 'address' then t2.value end) address,
  max(case when t2.`key` = 'city' then t2.value end) city,
  max(case when t2.`key` = 'region' then t2.value end) region,
  max(case when t2.`key` = 'country' then t2.value end) country,
  max(case when t2.`key` = 'postal_code' then t2.value end) postal_code,
  max(case when t2.`key` = 'phone' then t2.value end) phone
from table1 t1
left join table2 t2
  on t1.id = t2.id
group by t1.id, t1.name

See SQL Fiddle with Demo .

This could also be written using multiple joins on your table2 and you would include a filter on the join for each key :

select t1.id,
  t1.name,
  t2a.value address,
  t2c.value city,
  t2r.value region,
  t2y.value country,
  t2pc.value postal_code,
  t2p.value phone
from table1 t1
left join table2 t2a
  on t1.id = t2a.id
  and t2a.`key` = 'address'
left join table2 t2c
  on t1.id = t2c.id
  and t2c.`key` = 'city' 
left join table2 t2r
  on t1.id = t2r.id
  and t2c.`key` = 'region' 
left join table2 t2y
  on t1.id = t2y.id
  and t2c.`key` = 'country' 
left join table2 t2pc
  on t1.id = t2pc.id
  and t2pc.`key` = 'postal_code' 
left join table2 t2p
  on t1.id = t2p.id
  and t2p.`key` = 'phone';

See SQL Fiddle with Demo .

The above two versions will work great if you have a limited number of key values. If you have an unknown number of values, then you will want to look at using a prepared statement to generate dynamic SQL:

SET @sql = NULL;
SELECT
  GROUP_CONCAT(DISTINCT
    CONCAT(
      'max(case when t2.`key` = ''',
      `key`,
      ''' then t2.value end) AS `',
      `key`, '`'
    )
  ) INTO @sql
from Table2;

SET @sql 
    = CONCAT('SELECT t1.id, t1.name, ', @sql, ' 
              from table1 t1
              left join table2 t2
                on t1.id = t2.id
              group by t1.id, t1.name;');

PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;

See SQL Fiddle with Demo

All versions will give a result:

| ID | NAME |  ADDRESS |   CITY | REGION | COUNTRY | POSTAL_CODE |     PHONE |
|----|------|----------|--------|--------|---------|-------------|-----------|
|  1 |  Jim | X Street |     NY | (null) |  (null) |      (null) | 123456789 |
|  2 |  Bob |   (null) | (null) | (null) |  (null) |      (null) |    (null) |
|  3 | John |   (null) | (null) | (null) |  (null) |      (null) |    (null) |

Answer 2

You have a structure called entity-attribute-value in the second table. There are two ways to do the combination. I think the aggregation method is the easier to express:

select t1.name,
       max(case when `key` = 'address' then value end) as address,
       max(case when `key` = 'city' then value end) as city,
       max(case when `key` = 'region' then value end) as region,
       max(case when `key` = 'country' then value end) as country,
       max(case when `key` = 'postal_code' then value end) as postal_code,
       max(case when `key` = 'phone' then value end) as phone
from table1 t1 left join
     table2 t2
     on t1.id = t2.id
group by t1.name;

The second method is to do separate joins for each value:

select t1.name, address.value, city.value, . . .
from table1 t1 left join
     table2 address
     on t1.id = address.id and address.`key` = 'Address' left join
     table2 city
     on t1.id = city.id and city.`key` = 'City' . . .

Depending on the structure of the data, the join method can actually be faster in MySQL when it uses appropriate indexing. (Other databases have been algorithms for aggregation, so the group by method often works well in other databases.)