Count all items on which a user is the high bidder - mysql

Question

I am currently working on an auction system, and the functionality is all complete. I now need to add a count to the user's profile which shows how many items the user is currently bidding on.
The system comprises of two key tables (extra tables feature in the system of course, but these are the only tables related to this issue):
item_sales:
+-----+------------------+------------+-------------+---------+
| id | selling_format | duration | list_date | buyer |
+-----+------------------+------------+-------------+---------+
item_sales_bids:
+-----+-------------+-----------+---------------+-----------+--------+
| id | sale_item | user_id | current_bid | max_bid | date |
+-----+-------------+-----------+---------------+-----------+--------+
item_sales_bids.date is a Unix timestamp of the bid time.
I can easily get a count of all bids a given user has made with the following query:
SELECT COUNT(DISTINCT(`item_sales_bids`.`user_id`)) AS `total`,
SUM((`sale`.`list_date` + (`sale`.`duration` * 86400)) - UNIX_TIMESTAMP()) AS `endTime`
FROM `item_sales_bids`
INNER JOIN `item_sales` `sale` ON `item_sales_bids`.`sale_item` = `sale`.`id`
WHERE `user_id` = 1
GROUP BY `sale_item`
HAVING `endTime` > 0
What I would like to do, is run a query similar to the above, but only include records where the specified user is the current highest bidder (i.e. the max ID entry for a given item's bid set has a user_id value = to our user).
Unfortunately, I'm at a loss on how I might achieve this.
I have set up an SQLFiddle to assist > http://sqlfiddle.com/#!2/b98e4/3

Answer 1

Do a subs query to get the latest bid for all items and then join that to item_sales_bids to only process the latest items.
Something like this:-
SELECT COUNT(DISTINCT(item_sales_bids.user_id)) AS total,
SUM((sale.list_date + (sale.duration * 86400)) - UNIX_TIMESTAMP()) AS endTime
FROM item_sales_bids
INNER JOIN item_sales sale ON item_sales_bids.sale_item = sale.id
INNER JOIN
(
SELECT sale_item, MAX(id) AS LatestBid
FROM item_sales_bids
GROUP BY sale_item
) Sub1
ON item_sales_bids.sale_item = Sub1.sale_item AND item_sales_bids.id = Sub1.LatestBid
WHERE user_id = 1
GROUP BY item_sales_bids.sale_item
HAVING endTime > 0

Answer 2

This should do the trick. Toggling between user 1 & 2 will show the desired behavior. In the sample data, user 1 returns no data (they are not the high bidder) and user 2 returns a single row, with a current high bid of 50.
SELECT COUNT(DISTINCT(`bids`.`user_id`)) AS `total`,
`highbids`.`sale_item`,
`highbids`.`maxBid`,
SUM((`sale`.`list_date` + (`sale`.`duration` * 86400)) - UNIX_TIMESTAMP()) AS `endTime`
FROM `item_sales_bids` `bids`
INNER JOIN `item_sales` `sale` ON `bids`.`sale_item` = `sale`.`id`
INNER JOIN (SELECT MAX(`current_bid`) AS `maxBid`, `sale_item`, `user_id`
FROM `item_sales_bids`
GROUP BY `sale_item`, `user_id`
ORDER BY `current_bid` DESC LIMIT 1,1)
AS `highbids` ON `bids`.`user_id` = `highbids`.`user_id`
WHERE `bids`.`user_id` = 2
GROUP BY `bids`.`sale_item`
HAVING `endTime` > 0