I have a table where I keep messages and one where I keep users.
I want to get all the users that interactioned (send or received a message) with user_id 1.
This query works:
http://sqlfiddle.com/#!2/6a2f3/1
EDIT:
SELECT DISTINCT
(CASE WHEN `user_to_id` = 1 THEN `user_from_id` ELSE `user_to_id` END) `user_id`,
users.*
FROM `messages`
INNER JOIN users
ON (CASE WHEN `user_to_id` = 1 THEN `user_from_id` ELSE `user_to_id` END) = users.user_id
WHERE `user_to_id` = 1 OR `user_from_id` = 1
ORDER BY `time` DESC
But if I add to SELECT the message column, it returns duplicate records:
http://sqlfiddle.com/#!2/6a2f3/2
EDIT:
SELECT DISTINCT
(CASE WHEN `user_to_id` = 1 THEN `user_from_id` ELSE `user_to_id` END) `user_id`,
`messages`.`message`,
users.*
FROM `messages`
INNER JOIN users
ON (CASE WHEN `user_to_id` = 1 THEN `user_from_id` ELSE `user_to_id` END) = users.user_id
WHERE `user_to_id` = 1 OR `user_from_id` = 1
ORDER BY `time` DESC
How can I fix that?
And also, I see that it orders the results after the "DISTINCT" selection was made. The first query should return the results inverted because the row with message_id 2 has time 3.
Is there a way I can order them before the "DISTINCT"?
EDIT 2: I wasn't clear about the question. I want to select only the last message for a matched user_id.
Do you want something like this?
SELECT *
FROM (
SELECT
users.*,
(SELECT `message` from messages
WHERE
(CASE WHEN `user_to_id` = 1 THEN `user_from_id` ELSE `user_to_id` END) = users.user_id
AND (`user_to_id` = 1 OR `user_from_id` = 1)
ORDER BY `time` DESC limit 1
) AS message
FROM users
) a
WHERE message IS NOT NULL
SQL Fiddle
It's not returning duplicate records, you have two records with User_ID = 2.
I'm confused by what you want them to be ordered by. If you want to order them in the inverted order, just remove 'DESC'
Related
I am currently working on a project that has 2 very large sql tables Users and UserDocuments having around million and 2-3 millions records respectively. I have a query that will return the count of all the documents that each indvidual user has uploaded provided the document is not rejected.
A user can have multiple documents against his/her id.
My current query:-
SELECT
u.user_id,
u.name,
u.date_registered,
u.phone_no,
t1.docs_count,
t1.last_uploaded_on
FROM
Users u
JOIN(
SELECT user_id,
MAX(updated_at) AS last_uploaded_on,
SUM(CASE WHEN STATUS != 2 THEN 1 ELSE 0 END) AS docs_count
FROM
UserDocuments
WHERE
user_id IN(
SELECT
user_id
FROM
Users
WHERE
region_id = 1 AND city_id = 8 AND user_type = 1 AND user_suspended = 0 AND is_enabled = 1 AND verification_status = -1
) AND document_id IN('1', '2', '3', '4', '10', '11')
GROUP BY
user_id
ORDER BY
user_id ASC
) t1
ON
u.user_id = t1.user_id
WHERE
docs_count < 6 AND region_id = 1 AND city_id = 8 AND user_type = 1 AND user_suspended = 0 AND is_enabled = 1 AND verification_status = -1
LIMIT 1000, 100
Currently the query is taking very long around 20 secs to return data with indexes. can someone suggest some tweaks in the follwing query to gain some more preformance out of it.
SELECT
u.user_id,
max( u.name ) name,
max( u.date_registered ) date_registered,
max( phone_no ) phone_no,
MAX(d.updated_at) last_uploaded_on,
SUM(CASE WHEN d.STATUS != 2
THEN 1 ELSE 0 END) docs_count
FROM
Users u
JOIN UserDocuments d
ON u.user_id = d.user_id
AND d.document_id IN ('1', '2', '3', '4', '10', '11')
WHERE
u.region_id = 1
AND u.city_id = 8
AND u.user_type = 1
AND u.user_suspended = 0
AND u.is_enabled = 1
AND u.verification_status = -1
GROUP BY
u.user_id
HAVING
SUM(CASE WHEN d.STATUS != 2
THEN 1 ELSE 0 END) < 6
ORDER BY
u.user_id ASC
LIMIT
1000, 100
Have indexes on your tables as
user ( region_id, city_id, user_type, user_suspended, is_enabled, verification_status )
UserDocuments ( user_id, document_id, status, updated_at )
You are adding extra querying from the user table to both the inner and outer joins which might be killing it. Having an index on your critical "WHERE" components by user will pre-filter that set out. Only from that will it join to the UserDocuments table. By having the outer query get the counts() at the top level query.
Since the users name, registered and phone dont change per user, applying max() to each respectively prevents the need of adding those columns to the group by clause.
The index on the documents table on only the columns needed to confirm status and document_id and when last updated. This prevents the engine from having to go to the raw data pages as it can get the qualifying details directly from the index parts saving you time too.
LIMIT without ORDER BY does not make sense.
An ORDER BY in a 'derived table' is ignored.
Will you really have thousands of result rows? (I see the "offset of 1000".)
Use JOIN instead of IN ( SELECT ... )
What indexes do you have? I suggest INDEX(region_id, city_id, user_id)
CASE WHEN d.STATUS != 2 THEN 1 ELSE 0 END can be shortened to d.status != 2.
How many different values of status are there? If only two, then flip the test to d.status = 1`.
For my query, the two derived tables at the bottom are causing a crazy slow up for this query. The query, as is, takes about 45-55 seconds to execute.. NOW, when i remove just one of those derived tables (it does not matter which one) the query goes down to 0.1 - 0.3 seconds. My questions; Is there an issue with having multiple derived tables? Is there a better way to execute this? My indexes all seem to be correct, I will also include the explain from this query.
select t.name as team, u.name as "REP NAME",
count(distinct activity.id) as "TOTAL VISITS",
count(distinct activity.account_id) as "UNIQUE VISITS",
count(distinct placement.id) as "COMMITMENTS ADDED",
CASE WHEN
count(distinct activity.account_id) = 0 THEN (count(distinct
placement.id) / 1)
else (cast(count(distinct placement.id) as decimal(10,2)) /
cast(count(distinct activity.account_id) as decimal(10,2)))
end as "UNIQUE VISIT TO COMMITMENT %",
case when o.mode='basic' then count(distinct placement.id) else
count(distinct(case when placement.commitmentstatus='fullfilled'
then placement.id else 0 end))
end as "COMMITMENTS FULFILLED",
case when o.mode='basic' then 1 else
(CASE WHEN
count(distinct placement.id) = 0 THEN (count(distinct(case when
placement.commitmentstatus='fullfilled' then placement.id else 0
end)) / 1)
else (cast(count(distinct(case when
placement.commitmentstatus='fullfilled' then placement.id else 0
end)) as decimal(10,2)) / cast(count(distinct placement.id) as
decimal(10,2)))
end) end as "COMMITMENT TO FULFILLMENT %"
from lpmysqldb.users u
left join lpmysqldb.teams t on t.team_id=u.team_id
left join lpmysqldb.organizations o on o.id=t.org_id
left join (select * from lpmysqldb.activity where
org_id='555b918ae4b07b6ac5050852' and completed_at>='2018-05-01' and
completed_at<='2018-06-01' and tag='visit' and accountname is not
null and (status='active' or status='true' or status='1')) as
activity on activity.user_id=u.id
left join (select * from lpmysqldb.placements where
orgid='555b918ae4b07b6ac5050852' and placementdate>='2018-05-01' and
placementdate<='2018-06-01' and (status IN ('1','active','true') or
status is null)) as placement on placement.userid=u.id
where u.org_id='555b918ae4b07b6ac5050852'
and (u.status='active' or u.status='true' or u.status='1')
and istestuser!='1'
group by u.org_id, t.name, u.id, u.name, o.mode
order by count(distinct activity.id) desc
Thank you for assistance!
I have edited below with changing the two bottom joins from joining on subqueries to joining on the table directly. Still yielding the same result.
This is a SLIGHTLY restructured query of your same. Might be simplified as the last two subqueries are all pre-aggregated for your respective counts and count distincts so you can use those column names directly instead of showing all the count( distinct ) embedded throughout the query.
I also tried to simplify the division by multiplying a given count by 1.00 to force decimal-based precision as result.
select
t.name as team,
u.name as "REP NAME",
Activity.DistIdCnt as "TOTAL VISITS",
Activity.UniqAccountCnt as "UNIQUE VISITS",
Placement.DistIdCnt as "COMMITMENTS ADDED",
Placement.DistIdCnt /
CASE WHEN Activity.UniqAccountCnt = 0
THEN 1.00
ELSE Activity.UniqAccountCnt * 1.00
end as "UNIQUE VISIT TO COMMITMENT %",
case when o.mode = 'basic'
then Placement.DistIdCnt
else Placement.DistFulfillCnt
end as "COMMITMENTS FULFILLED",
case when o.mode = 'basic'
then 1
else ( Placement.DistFulfillCnt /
CASE when Placement.DistIdCnt = 0
then 1.00
ELSE Placement.DistIdCnt * 1.00
END TRANSACTION )
END as "COMMITMENT TO FULFILLMENT %"
from
lpmysqldb.users u
left join lpmysqldb.teams t
on u.team_id = t.team_id
left join lpmysqldb.organizations o
on t.org_id = o.id
left join
( select
user_id,
count(*) as AllRecs,
count( distinct id ) DistIdCnt,
count( distinct account_id) as UniqAccountCnt
from
lpmysqldb.activity
where
org_id = '555b918ae4b07b6ac5050852'
and completed_at>='2018-05-01'
and completed_at<='2018-06-01'
and tag='visit'
and accountname is not null
and status IN ( '1', 'active', 'true')
group by
user_id ) activity
on u.id = activity.user_id
left join
( select
userid,
count(*) AllRecs,
count(distinct id) as DistIdCnt,
count(distinct( case when commitmentstatus = 'fullfilled'
then id
else 0 end )) DistFulfillCnt
from
lpmysqldb.placements
where
orgid = '555b918ae4b07b6ac5050852'
and placementdate >= '2018-05-01'
and placementdate <= '2018-06-01'
and ( status is null OR status IN ('1','active','true')
group by
userid ) as placement
on u.id = placement.userid
where
u.org_id = '555b918ae4b07b6ac5050852'
and u.status IN ( 'active', 'true', '1')
and istestuser != '1'
group by
u.org_id,
t.name,
u.id,
u.name,
o.mode
order by
activity.DistIdCnt desc
FINALLY, your inner queries are querying for ALL users. If you have a large count of users that are NOT active, you MIGHT exclude those users from each inner query by adding those join/criteria there too such as...
( ...
from
lpmysqldb.placements
JOIN lpmysqldb.users u2
on placements.userid = u2.id
and u2.status IN ( 'active', 'true', '1')
and u2.istestuser != '1'
where … ) as placement
I want to get results for likes by a specific user-id and all likes and dislikes of a specific page id.
My structure looks like this:
`pages`: (id, title)
`pages_likes`: (id, page_id, uid, status)
If the status is -1 it's a dislike of a specific page, if it's 1 it's a like.
So, to get all likes this is my query:
SELECT COUNT(id) FROM pages_likes WHERE status = '1'
But now I also want to get if user-id 3 for example likes this page with
SELECT COUNT(id) FROM pages_likes WHERE status = '1' AND uid='3'
How can I achieve both in one query? I guess there has to be changed something right after the SELECT statement?
If you want to do this in a single query, use conditional aggregation:
SELECT SUM(CASE WHEN uid = '3' THEN 1 ELSE 0 END) AS threeLikes,
SUM(CASE WHEN uid <> '3' THEN 1 ELSE 0 END) AS otherLikes
FROM pages_likes
WHERE status = '1'
Another option would be to use a UNION, cf. the answer given by #bernie
Update:
If you want page likes and dislikes in the same query, you can try:
SELECT SUM(CASE WHEN uid = '3' AND status = '1' THEN 1 ELSE 0 END) AS threeLikes,
SUM(CASE WHEN uid <> '3' AND status = '1' THEN 1 ELSE 0 END) AS otherLikes
SUM(CASE WHEN uid = '3' AND status = '0' THEN 1 ELSE 0 END) AS threeDislikes,
SUM(CASE WHEN uid <> '3' AND status = '0' THEN 1 ELSE 0 END) AS otherDisikes
FROM pages_likes
I think this would work:
SELECT '' user, COUNT(id) likes
FROM pages_likes
WHERE status = '1'
GROUP BY 1
UNION ALL
SELECT uid user, COUNT(id) likes
FROM pages_likes
WHERE status = '1' AND uid='3'
GROUP BY 1
I would do it this way:
SELECT COUNT(*) AS total_likes, SUM(uid=3) AS uid3_likes
FROM pages_likes
WHERE status=1 AND page_id=1234
Re your comment:
Here's an example of showing total likes, and total dislikes. It's similar to the answer from #TimBiegeleisen.
SELECT SUM(status=1) AS total_likes,
SUM(CASE WHEN uid=3 AND status=1 THEN 1 END) AS uid3_likes
SUM(CASE WHEN uid=3 AND status=-1 THEN 1 END) AS uid3_dislikes
FROM pages_likes
WHERE page_id=1234
I have 2 tables odds and matches :
matches : has match_id and match_date
odds : has id, timestamp, result, odd_value, user_id, match_id
I had a query that get the following information from those tables for each user:
winnings : the winning bets for each user. (when odds.result = 1)
loses : the lost bets for each user.(when odds.result != 1)
points : the points of each user.(the sum of the odds.odd_value) for each user.
bonus : for each continuous 5 winnings i want to add extra bonus to this variable. (for each user)
How to calculate bonus?
I tried to use this query and I faced a problem : (you can check it here SQL Fiddle)
the calculated bonus are not right for all the users :
first user:(winnings:13, bonus=2).
second user:(winnings:8, bonus=2)bonus here should be 1.
third user:(winnings:14, bonus=3)bonus here should be 2.
why does the query not calculate the bonus correctly?
select d.user_id,
sum(case when d.result = 1 then 1 else 0 end) as winnings,
sum(case when d.result = 2 then 1 else 0 end) as loses,
sum(case when d.result = 1 then d.odd_value else 0 end) as points,
f.bonus
FROM odds d
INNER JOIN
(
SELECT
user_id,SUM(CASE WHEN F1=5 THEN 1 ELSE 0 END) AS bonus
FROM
(
SELECT
user_id,
CASE WHEN result=1 and #counter<5 THEN #counter:=#counter+1 WHEN result=1 and #counter=5 THEN #counter:=1 ELSE #counter:=0 END AS F1
FROM odds o
cross join (SELECT #counter:=0) AS t
INNER JOIN matches mc on mc.match_id = o.match_id
WHERE MONTH(STR_TO_DATE(mc.match_date, '%Y-%m-%d')) = 2 AND
YEAR(STR_TO_DATE(mc.match_date, '%Y-%m-%d')) = 2015 AND
(YEAR(o.timestamp)=2015 AND MONTH(o.timestamp) = 02)
) Temp
group by user_id
)as f on f.user_id = d.user_id
group by d.user_id
I am not sure how your result related to matches table,
you can add back WHERE / INNER JOIN clause if you need.
Here is link to fiddle
and the last iteration according to your comments:
And here is a query:
SET #user:=0;
select d.user_id,
sum(case when d.result = 1 then 1 else 0 end) as winnings,
sum(case when d.result = 2 then 1 else 0 end) as loses,
sum(case when d.result = 1 then d.odd_value else 0 end) as points,
f.bonus
FROM odds d
INNER JOIN
(
SELECT
user_id,SUM(bonus) AS bonus
FROM
(
SELECT
user_id,
CASE WHEN result=1 and #counter<5 AND #user=user_id THEN #counter:=#counter+1
WHEN result=1 and #counter=5 AND #user=user_id THEN #counter:=1
WHEN result=1 and #user<>user_id THEN #counter:=1
ELSE
#counter:=0
END AS F1,
#user:=user_id,
CASE WHEN #counter=5 THEN 1 ELSE 0 END AS bonus
FROM odds o
ORDER BY user_id , match_id
) Temp
group by user_id
)as f on f.user_id = d.user_id
group by d.user_id
This is what I have at the moment.
$db =& JFactory::getDBO();
$query = $db->getQuery(true);
$query->select('`#__catalog_commit`.`id` as id, `#__catalog_commit`.`date` as date, COUNT(`#__catalog_commit_message`.`commit_id`) as count,
(SELECT COUNT(`#__catalog_commit_message`.`type`) as count_notice FROM `#__catalog_commit_message` WHERE `#__catalog_commit_message`.`type` = 1 GROUP BY `#__catalog_commit_message`.`type`) as count_notice,
(SELECT COUNT(`#__catalog_commit_message`.`type`) as count_warning FROM `#__catalog_commit_message` WHERE `#__catalog_commit_message`.`type` = 2 GROUP BY `#__catalog_commit_message`.`type`) as count_warning,
(SELECT COUNT(`#__catalog_commit_message`.`type`) as count_error FROM `#__catalog_commit_message` WHERE `#__catalog_commit_message`.`type` = 3 GROUP BY `#__catalog_commit_message`.`type`) as count_error');
$query->from('#__catalog_commit_message');
$query->leftjoin('`#__catalog_commit` ON `#__catalog_commit`.`id` = `#__catalog_commit_message`.`commit_id`');
$query->group('`#__catalog_commit_message`.`commit_id`');
$query->order('`#__catalog_commit`.`id` DESC');
What I have is 2 tables with the following structures:
catalog_commit
==============
id
date
catalog_commit_message
======================
id
commit_id
type
message
Basically I want to have the count of each different types of messages per group items. In what I have it actually select every rows (Which is normal) but I'm looking for a way (nicier if possible) to have the count per messages type within the query.
EDIT: Just wanted to add that it's a JModelList.
From what I gather, this should be your query:
SELECT c.id
,c.date
,count(cm.commit_id) as ct_total
,sum(CASE WHEN cm.type = 1 THEN 1 ELSE 0 END) AS count_notice
,sum(CASE WHEN cm.type = 2 THEN 1 ELSE 0 END) AS count_warning
,sum(CASE WHEN cm.type = 3 THEN 1 ELSE 0 END) AS count_error
FROM catalog_commit c
LEFT JOIN catalog_commit_message cm ON cm.commit_id = c.id
GROUP BY c.id, c.date
ORDER BY c.id DESC
You had the order of your tables reversed in the LEFT JOIN. Also, you had weird subqueries in the SELECT list.