Question

I have a setup for 2 tables which is rather curious.

Table A:
it has multiple columns, but is basically a collection of information.
Say:

------------------------------------
my_id| name | address | lat | long 

Table B is a long list of relationshisp between elements of table A:
For example:

------------------
my_id | parent_id

What I want to do is the following:

For any A.my_id, get the list of the B.my_id which have A.my_id as B.parent_id (there is also the case in which there’s noone) and then:

get, for each element in this list (we just got), lats and longs from table A for these ids and create an 2 array out of them.

Can anyone point me in the right direction? I am rather new to SQL and while I managed up ’til now, this really baffles me.

Thanks!

Answer 1

SELECT results.lat, results.long 
FROM a 
JOIN b 
    ON b.parent_id=a.my_id 
JOIN a AS results 
    ON results.my_id=b.my_id
WHERE a.my_id = 'the one I'm looking for'

Ignoring the selected columns for a minute, the query will give you rows from table A that match the particular ID you’re looking for (see the where clause), joined with rows from B that have the same parent ID as the ID you’re looking for, joined with table A again, this time rows that have an ID that matches the ID in B.

If your tables look like this:

TABLE A
my_id| name | address | lat | long
  1    ABC    1234      10     20
  2    DEF    5678      15     19
  3    GHI    9101      17     21
  4    JKL    1121      18     18
 ...

TABLE B
my_id|parent_id
  1      4
  3      4
 ...

If you do this query:

SELECT *
FROM a 
JOIN b 
    ON b.parent_id=a.my_id 
JOIN a AS results 
    ON results.my_id=b.my_id
WHERE a.my_id = 4

You should get these results:

a.my_id | a.name | a.address | a.lat | a.long | b.my_id | b.parent_id | results.my_id | results.name | results.address | results.lat | results.long
   4       JKL       1121       18       18        1          4             1             ABC                   1234           10             20
   4       JKL       1121       18       18        3          4             3             GHI                   9101           17             21

From there we just change the columns in the SELECT to affect what the results look like.

The trick here is using the same table (A) in two places in the JOIN.