Question

I have table 1 that represents Items with column ItemID and column ItemInfo.
Then, I have table 2 that represents SubItems with column ItemID (the parent), column SubItemID and column SubItemInfo.

I’m trying to get the count of all subitems for each item with this :

select items.itemID, count(*) as count 
from items, subItems 
where subItems.itemID=items.itemID group by itemID

It works fine except for items that don’t have sub-items. Instead of returning the item row with a sub-item count of 0, the row is simply not there.

Is there an efficient way to force all rows from the first table (Items) to be selected ?

Answer 1

You neet to use left outer join, than you’ll get 0 for items having no subitems, like that:

SELECT items.itemid,
       Count(subitems.subitemid) AS count
FROM   items
       LEFT JOIN subitems
              ON ( subitems.itemid = items.itemid )
GROUP  BY items.itemid

If you wanted to have better performance, you could try this query:

SELECT itemid,
       Count(*) AS count
FROM   subitems
GROUP  BY itemid

BUT, with this one you will not get any info about the items, that have no subitems (they are simply not present within subitems table).