问题
My table contains votes of users for different items. It has the following fields:
id, user_id, item_id, vote, utc_time
I understand how to get the last vote of #user# for #item#, but it uses subquery:
SELECT votes.*, items.name, items.price
FROM votes JOIN items ON items.id = votes.item_id
WHERE user_id = #user# AND item_id = #item#
AND utc_time = (
SELECT MAX(utc_time) FROM votes
WHERE user_id = #user# AND item_id = #item#
)
It works, but it looks quite stupid to me... There should be a more elegant way to get this one record. I tried the approach suggested here, but I cannot make it work yet, so I'll appreciate your help: How can I SELECT rows with MAX(Column value), DISTINCT by another column in SQL?
There is a second part to this question: Count rows with DISTINCT(several columns) and MAX(another column)
回答1:
You want just one row from the result, the one with MAX(utc_time)
. In MySQL, there is a LIMIT
clause you can apply with ORDER BY
:
SELECT votes.*, items.name, items.price
FROM votes JOIN items ON items.id = votes.item_id
WHERE user_id = #user# AND item_id = #item#
ORDER BY votes.utc_time DESC
LIMIT 1 ;
An index on either (user_id, item_id, utc_time)
or (item_id, user_id, utc_time)
will be good for efficiency.
回答2:
Simple: if the date/time is the maximum date there will not exist a "higher" (more recent) date/time (for the same {user,item} ).
SELECT vo.*
, it.name, it.price
FROM votes vo
JOIN items it ON it.id = vo.item_id
WHERE vo.user_id = #user# AND vo.item_id = #item#
AND NOT EXISTS (
SELECT *
FROM votes nx
WHERE nx.user_id = vo.user_id
AND nx.item_id = vo.item_id
AND nx.utc_time > vo.utc_time
);
来源:https://stackoverflow.com/questions/14430260/select-one-row-with-maxcolumn-for-known-other-several-columns-without-subquery