Question

Table: Unit
ID NAME  VALUE
1  Kilo  1000
2  Mega  1000000
3  Giga  1000000000

Table: Storage
ID Title   Drive_value DriveUnitID Cache_value CacheUnitID Status_ID error error_unit area
1  Seagate  100         3          400         1           2         1     1          1
2  Scansoft 250         3          80          2           1         1     2          2

Table: manufac
    ID   area 
    1    US
    2    CHINA 

Table: Status
ID   Description 
1    Blah.. Blah
2    Durka Durka

Desired goal is to have a subselect joining three tables
Something like this (below) but this syntax doesnt work

select 
s.title, 
t.description,
x.area,
u1.value * s.cache_value as Cache,
u2.value * s.drive_value as Drive,
u3.value * s.error_value as Error
((u4.value * s.error_value)+(u4.value * s.error_value)) as ErrorHigh
((u5.value * s.error_value)-(u5.value * s.error_value)) as ErrorHigh
from storage s
join status t on t.id = s.status_id
join manufac x on x.id = s.manufac_id 
join unit u1 on s.cache_unit_id = u1.id
join unit u2 on s.drive_unit_id = u2.id
join unit u3 on s.error_unit_id = u3.id
join unit u4 on s.error_unit_id = u4.id
join unit u5 on s.error_unit_id = u5.id

Answer 1

This queries are semantically the same. But remember either you put the table in the join or in the from clause:

With explicit joins:

select s.title, t.description,
u1.value * s.cache_value as Cache,
u2.value * s.drive_value as Drive
from storage s
join status t on t.id = s.status_id
join unit u1 on s.cache_unit_id = u1.id
join unit u2 on s.drive_unit_id = u2.id

With implicit joins:

select s.title, t.description,
u1.value * s.cache_value as Cache,
u2.value * s.drive_value as Drive
from storage s, status t, unit u1, unit u2
where t.id = s.status_id and s.cache_unit_id = u1.id and s.drive_unit_id = u2.id

The final result is the same for the end user and the DBMS also process both query the same way, so there are no performance increment by using one or the other.

Edit:

After the requirements changed this is the query I think you’re looking for:

select s.title, t.description,
u1.value * s.cache_value as Cache,
u2.value * s.drive_value as Drive,
u3.value * s.error_value as Error,
u3.value * s.error_value * 2 as ErrorHigh,
u3.value * s.error_value - u3.value * s.error_value as ErrorLow,
m.area
from storage s
join status t on t.id = s.status_id
join manufac m on m.id = s.area
join unit u1 on s.cache_unit_id = u1.id
join unit u2 on s.drive_unit_id = u2.id
join unit u3 on s.error_unit_id = u3.id

Example

Result:

+----------+-------------+----------+--------------+---------+-----------+----------+-------+
|  TITLE   | DESCRIPTION |  CACHE   |    DRIVE     |  ERROR  | ERRORHIGH | ERRORLOW | AREA  |
+----------+-------------+----------+--------------+---------+-----------+----------+-------+
| Seagate  | Durka Durka |   400000 | 100000000000 |    1000 |      2000 |        0 | US    |
| Scansoft | Blah.. Blah | 80000000 | 250000000000 | 1000000 |   2000000 |        0 | CHINA |
+----------+-------------+----------+--------------+---------+-----------+----------+-------+

• Notice in your question there are two columns named ErrorHigh.
• Math “errors” x + x = 2 * x (where x = u4.value * s.error_value)
• Probably the worst math error 🙂 x – x = 0 (where x = u5.value * s.error_value)

So I guess this query is not what you really need, but should be enough to lead you in the right direction. Good luck!