避免多次连接同一个表

Question

I need help in avoiding joining same table multiple times. Here is my sample query.

Table1 have 3 columns ABC_ID,DEF_ID,XYZ_ID and is associated to table2 on ID column . 3 column values in table1 can be null and If values are present then i need to return associated value from table 2 using different column names as in select statement.

As i am using left join i ended up joining table2 three times with table1 on each type if ID column which is having lot of performance issues. How can i write this in different way to avoid join multiple times. Here is sample data. Any help is appreciated

    select 
          (
          CASE
            WHEN ( table2.ID = table1.ABC_ID)
            THEN table2.ID_VAL
            ELSE 'TEST1'
          END ) AS "TEST1",
          (
          CASE
            WHEN (table2a.ID = table1.DEF_ID)
            THEN table2a.ID_VAL
            ELSE 'TEST2'
          END ) AS "TEST2",
          (
          CASE
            WHEN (table2b.ID = table1.XYZ_ID)
            THEN table2b.ID_VAL
            ELSE 'TEST3'
          END ) AS "TEST3"
    from table1 table1 
    left join table2 table2 on   (table2.ID=table1.ABC_ID)
    left join table2 table2a on   ( table2a.id=table1.DEF_ID)
    left join table2 table2b on   ( table2b.id=table1.XYZ_ID)
    where table1.Id_NUM='1'

Table1

    Id_NUM  ABC_ID     DEF_ID    XYZ_ID
    1       12345      456789    32145
    2       null       456789    32145
    3       12345      null      null

Table2

    ID         ID_VAL
    12345      abcded
    456789     kjwsddk
    321456     wedfgfv

OUTPUT

    TEST1         TEST2       TEST3
    12345         456789      32145

Answer 1

Your join s are fine, but the query can be simplified:

select coalesce(t2a.ID_VAL, 'TEST1') as test1,
       coalesce(t2d.ID_VAL, 'TEST2') as test2,
       coalesce(t2x.ID_VAL, 'TEST3') as test3
from table1 t1 left join
     table2 t2a
     on t2a.ID = t1.ABC_ID left join
     table2 t2d
     on t2d.id = t1.DEF_ID left join
     table2 t2x
     on t2x.id = t1.XYZ_ID
where t1.Id_NUM = 1;

Notes:

• The table aliases should be meaningful. In this case, the aliases include abbreviations for the column used for the join.
• Presumably, id_num is a number, so don't use single quotes for the value.
• There is no reason to escape the column names. The double quotes just add query clutter.
• The case expressions can be replaced by coalesce() , which is simpler (this is not exactly the same if id_val can be NULL , but I'm guessing that doesn't happen).

There is no reason to avoid joining the table three times. That is what your data model requires, because you have three separate foreign key references.