SQL LEFT JOIN和WHERE产生意外结果

Question

I've got two tables related by a field called 'status'. With table2.location having a value of 'texas', there are 20 instances of table2.status having a value of 'good'.

That is, the following query returns 20 rows of table2.status = 'good'.

[A] SELECT table2.status FROM table2 WHERE table2.location = 'texas';

Further, there are 50 unique table1.id's with table1.status = 'good'.

That is, the following query returns 50 rows of unique table1.id's.

[B] SELECT table1.id FROM table1 WHERE table1.status = 'good';

Now, when I run the following query:

[C] SELECT table1.id FROM table1 LEFT JOIN table2 ON table1.status = table2.status WHERE table2.location = "texas";

I would expect it to return the 50 rows of unique id's. However, it is actually returning the 50 rows of unique id's 20 times (ie. I get 1000 rows returned).

The quick fix I did was to simply execute SELECT DISTINCT table1.id... which then just returns one set of 50 rows (not 20 sets of 50 rows).

However, I'd like to know why I'm seeing this behahvior - maybe there is something wrong with my query [C] ?

Thanks!

Answer 1

Your query is wrong if you want to get 50 items: you're performing a cross join 50x20, so you have 1000 records as result.
You cannot join on status (which is not unique): probably your table design is wrong.
IMO you should have id in both tables and join with it...

Answer 2

it is expected. 1st row from table 1 matches 50 rows from table 2, 2nd row from table 1 matches again the same 50 rows from table 2 etc.

Answer 3

Your query has two problems

SELECT table1.id 
FROM table1 
LEFT JOIN table2 ON table1.status = table2.status 
WHERE table2.location = "texas";

First when you usae a left join a condition on the second table must be in the ON clause not the where clause or you convert the left join to an inner join (Since the condition must be met)

So your query should start looking like this:

SELECT table1.id 
FROM table1 
LEFT JOIN table2 
    ON table1.status = table2.status 
    AND table2.location = "texas";

Now your next problem is that status is unlikely to be the thing you actually want to join on. To help you get the results you want though, we would need to see the table structure of the two tables.