How to do a IFNULL(count(*),0) in MySQL 5.5 - mysql

I tried to do a IFNULL(count(),0) , or IF (count() > 0, count(*),0),
but it doesn't work, my row "counter" display "NULL" instead of 0 :/
Here is my query:
SELECT IF(fc.counter > 0, fc.counter, 0) counter, b.*, fc.* FROM client_branche cb INNER JOIN branche b On b.id = cb.branche_id LEFT OUTER JOIN (
SELECT count(*) as counter, ctn_b.branche_id as b_id
FROM `historique` h
INNER JOIN contenu_branche ctn_b ON ctn_b.contenu_id = h.contenu_id
INNER JOIN utilisateur u ON u.id = h.utilisateur_id
WHERE h.h_fini = 1 AND ( u.client_id = 1 OR u.client_id = 0 ) AND h.h_dateheure BETWEEN '2015-12-24' AND '2015-12-30'
group by ctn_b.`branche_id`)
fc ON fc.b_id = cb.branche_id WHERE cb.client_id = 1
So i tried to do that :
SELECT IF(fc.counter > 0, fc.counter, 0) counter, b.*, fc.* FROM client_branche cb INNER JOIN branche b On b.id = cb.branche_id LEFT OUTER JOIN (
SELECT IFNULL(count(*),0) as counter, ctn_b.branche_id as b_id
FROM `historique` h
INNER JOIN contenu_branche ctn_b ON ctn_b.contenu_id = h.contenu_id
INNER JOIN utilisateur u ON u.id = h.utilisateur_id
WHERE h.h_fini = 1 AND ( u.client_id = 1 OR u.client_id = 0 ) AND h.h_dateheure BETWEEN '2015-12-24' AND '2015-12-30'
group by ctn_b.`branche_id`)
fc ON fc.b_id = cb.branche_id WHERE cb.client_id = 1
Annnnd i failed. I hope someone will help me. Thanks a lot in advance, and sorry for my bad english, it's not my native language :)

The problem is probably that you have is multiple columns in the select have the same name -- counter. The best way to solve this is not to use * in the query, but to list the columns you want. Perhaps a simpler way is to rename the column:
SELECT (CASE WHEN fc.counter > 0 THEN fc.counter ELSE 0 END) as fc_counter,
b.*, fc.*
FROM client_branche cb INNER JOIN
branche b
On b.id = cb.branche_id LEFT OUTER JOIN
(SELECT count(*) as counter, ctn_b.branche_id as b_id
FROM `historique` h INNER JOIN
contenu_branche ctn_b
ON ctn_b.contenu_id = h.contenu_id INNER JOIN
utilisateur u
ON u.id = h.utilisateur_id
WHERE h.h_fini = 1 AND
u.client_id IN (0, 1) AND
h.h_dateheure BETWEEN '2015-12-24' AND '2015-12-30'
GROUP BY ctn_b.branche_id
) fc
ON fc.b_id = cb.branche_id
WHERE cb.client_id = 1;
I made a couple other small changes as well, such as using IN instead of OR and CASE (ANSI standard) instead of IF().
Note: If counter is never negative, the most colloquial way of writing the logic is COALESCE(counter, 0) rather than CASE or IF().

It will work. You are use counter alias in subquery and also main query. So try it
SELECT IF(fc.cou > 0, fc.cou, 0) counter, b., fc.FROM client_branche cb INNER JOIN branche b On b.id = cb.branche_id LEFT OUTER JOIN ( SELECT IFNULL(count(*),0) as cou, ctn_b.branche_id as b_id FROM historiqueh INNER JOIN contenu_branche ctn_b ON ctn_b.contenu_id = h.contenu_id INNER JOIN utilisateur u ON u.id = h.utilisateur_id WHERE h.h_fini = 1 AND ( u.client_id = 1 OR u.client_id = 0 ) AND h.h_dateheure BETWEEN '2015-12-24' AND '2015-12-30' group by ctn_b.branche_id) fc ON fc.b_id = cb.branche_id WHERE cb.client_id = 1

You need the IFNULL on the outer query (due to the LEFT JOIN), like so:
SELECT IF(IFNULL(fc.counter,0) > 0, IFNULL(fc.counter,0), 0) AS counter
, b.*, fc.*
FROM client_branche AS cb
INNER JOIN branche AS b ON b.id = cb.branche_id
LEFT JOIN (
SELECT count(*) AS counter, ctn_b.branche_id AS b_id
FROM `historique` AS h
INNER JOIN contenu_branche AS ctn_b ON ctn_b.contenu_id = h.contenu_id
INNER JOIN utilisateur AS u ON u.id = h.utilisateur_id
WHERE h.h_fini = 1 AND ( u.client_id = 1 OR u.client_id = 0 )
AND h.h_dateheure BETWEEN '2015-12-24' AND '2015-12-30'
GROUP BY ctn_b.`branche_id`
) AS fc ON fc.b_id = cb.branche_id
WHERE cb.client_id = 1
;
It is effectively a "no op" on the inner query, since count(*) there will never be null.

Related

MYSQL combining two counts that use same columns

I have two queries that i need to combine into one query. The problem is they are using the same columns in where clause depending on what i need to fetch.
Query 1
SELECT c.fullname, COUNT( DISTINCT sst.id ) AS 'Liczba rozpoczetych szkolen'
FROM mdl_course c
INNER JOIN mdl_scorm s ON s.course = c.id
INNER JOIN mdl_scorm_scoes_track sst ON s.id = sst.scormid
INNER JOIN mdl_user u ON u.id = sst.userid
WHERE sst.element = 'x.start.time' AND u.deleted =0
GROUP BY c.fullname ORDER BY `Liczba rozpoczetych szkolen` ASC
Query 2
SELECT c.fullname, COUNT(DISTINCT sst.userid ) AS 'Liczba_ukonczonych_szkolen'
FROM mdl_course c
INNER JOIN mdl_scorm s ON s.course=c.id
INNER JOIN mdl_scorm_scoes_track sst ON s.id = sst.scormid
INNER JOIN mdl_user u ON sst.userid=u.id
where `element`='cmi.core.score.raw' and `value` = 100 and u.deleted = 0
GROUP BY c.fullname ORDER BY `Liczba_ukonczonych_szkolen` DESC
They are depending on the same column named 'element'.
How i can display the result as
fullname Liczba rozpoczetych szkolen Liczba_ukonczonych_szkolen
A1 34 4
A2 5 3
A3 34 33
I've came up with this one, whitch works. Thanks for advice #HoneyBadger
SELECT t1.fullname, t1.Liczba_rozpoczetych_szkolen, t2.Liczba_ukonczonych_szkolen,
round(((t2.Liczba_ukonczonych_szkolen /t1.Liczba_rozpoczetych_szkolen)*100),2) as procentowo
FROM
(SELECT c.fullname, COUNT( DISTINCT sst.id ) AS Liczba_rozpoczetych_szkolen
FROM mdl_course c
INNER JOIN mdl_scorm s ON s.course = c.id
INNER JOIN mdl_scorm_scoes_track sst ON s.id = sst.scormid
INNER JOIN mdl_user u ON u.id = sst.userid
WHERE sst.element = 'x.start.time' AND u.deleted =0
GROUP BY c.fullname ) AS t1,
(SELECT c.fullname, COUNT(DISTINCT sst.userid ) AS Liczba_ukonczonych_szkolen
FROM mdl_course c
INNER JOIN mdl_scorm s ON s.course=c.id
INNER JOIN mdl_scorm_scoes_track sst ON s.id = sst.scormid
left join mdl_user u ON sst.userid=u.id
where `element`='cmi.core.score.raw' and `value` = 100 and u.deleted = 0
GROUP BY c.fullname) as t2
WHERE t1.fullname = t2.fullname
You can join them together like this:
SELECT COALESCE(R.fullname, U.fullname) AS fullname
, COALESCE(R.Liczba_rozpoczetych_szkolen, 0) AS Liczba_rozpoczetych_szkolen
, COALESCE(R.Liczba_ukonczonych_szkolen, 0) AS Liczba_ukonczonych_szkolen
FROM (
SELECT c.fullname
, COUNT( DISTINCT sst.id ) AS Liczba_rozpoczetych_szkolen
FROM mdl_course c
INNER JOIN mdl_scorm s
ON s.course = c.id
INNER JOIN mdl_scorm_scoes_track sst
ON s.id = sst.scormid
INNER JOIN mdl_user u
ON u.id = sst.userid
WHERE sst.element = 'x.start.time'
AND u.deleted =0
GROUP BY c.fullname
) AS R
FULL JOIN (
SELECT c.fullname
, COUNT(DISTINCT sst.userid ) AS Liczba_ukonczonych_szkolen
FROM mdl_course c
INNER JOIN mdl_scorm s
ON s.course=c.id
INNER JOIN mdl_scorm_scoes_track sst
ON s.id = sst.scormid
INNER JOIN mdl_user u
ON sst.userid=u.id
where `element`='cmi.core.score.raw'
AND `value` = 100
AND u.deleted = 0
GROUP BY c.fullname
) AS U
ON R.fullname = U.fullname
Try the query with the ORed predicates and extra filtering with CASE in the SELECT list. Kind of
SELECT c.fullname, COUNT(DISTINCT CASE WHEN sst.element = 'x.start.time' THEN sst.id END) AS 'Liczba rozpoczetych szkolen'
, COUNT(DISTINCT CASE WHEN sst.element = 'cmi.core.score.raw' and `value` = 100 THEN sst.userid END) AS 'Liczba_ukonczonych_szkolen'
FROM mdl_course c
INNER JOIN mdl_scorm s ON s.course = c.id
INNER JOIN mdl_scorm_scoes_track sst ON s.id = sst.scormid
INNER JOIN mdl_user u ON u.id = sst.userid
WHERE sst.element = IN( 'x.start.time', 'cmi.core.score.raw') AND u.deleted =0
GROUP BY c.fullname ORDER BY `Liczba rozpoczetych szkolen` ASC

Select threads from forum based on comments

I've got the following sql that will return a list of forums. Under each forum it will select the thread with the latest comment. This works fine but when a new thread hasn't got any comments, nothing is returned.
How to tackle this problem?
SELECT f.Id AS forum_id,
f.name AS forum_name,
f.slug AS forum_slug,
f.image AS forum_image,
t.Id AS thread_id,
t.title AS thread_topic,
t.unixtime AS thread_timestamp,
p.Id AS post_id,
p.content AS post_content,
p.unixtime AS post_timestamp,
(SELECT COUNT(*) FROM a_comments o WHERE o.forumID=f.Id AND o.teamId = {$teamId}) comments_count,
(SELECT COUNT(*) FROM a_threads w WHERE w.forumID=f.Id AND w.teamId = {$teamId}) threads_count
FROM a_forums f
LEFT JOIN (SELECT t2.forumID, max(COALESCE(p2.unixtime, t2.unixtime)) as ts, COUNT(p2.unixtime) as post_count
FROM a_threads t2
LEFT JOIN a_comments p2 ON p2.threadId = t2.id
GROUP BY t2.forumId) max_p ON f.id = max_p.forumId
LEFT JOIN a_comments p ON max_p.ts = p.unixtime AND p.teamId = {$teamId} AND p.deleted = 0
LEFT JOIN a_threads t ON f.Id = t.forumID AND (max_p.post_count = 0 OR p.threadId = t.ID) AND t.teamId = {$teamId} AND t.deleted = 0
ORDER BY f.id
I think you just have to change the LEFT JOIN in the first subquery to a JOIN. With the LEFT JOIN, you'll get NULL or a non-valid time for the comment. This then throws off the rest of the logic -- I think.
SELECT f.Id AS forum_id, f.name AS forum_name, f.slug AS forum_slug, f.image AS forum_image,
t.Id AS thread_id, t.title AS thread_topic, t.unixtime AS thread_timestamp,
p.Id AS post_id, p.content AS post_content, p.unixtime AS post_timestamp,
(SELECT COUNT(*) FROM a_comments o WHERE o.forumID=f.Id AND o.teamId = {$teamId}) as comments_count,
(SELECT COUNT(*) FROM a_threads w WHERE w.forumID=f.Id AND w.teamId = {$teamId}) as threads_count
FROM a_forums f LEFT JOIN
(SELECT t2.forumID, max(p2.unixtime) as ts,
COUNT(p2.unixtime) as post_count
FROM a_threads t2 JOIN
a_comments p2
ON p2.threadId = t2.id
GROUP BY t2.forumId
) max_p
ON f.id = max_p.forumId LEFT JOIN
a_comments p
ON max_p.ts = p.unixtime AND p.teamId = {$teamId} AND
p.deleted = 0 LEFT JOIN
a_threads t
ON f.Id = t.forumID AND (max_p.post_count = 0 OR p.threadId = t.ID) AND t.teamId = {$teamId} AND t.deleted = 0
ORDER BY f.id

How do select row with max(column value) group by another column in SQL?

i dont't know how to select row with max column value group by another column. I have T-SQL
CREATE PROC GET_USER
AS
BEGIN
SELECT A.USER_ID, B.START_DATE , D.START_DATE, A.FULL_NAME,A.COST_CENTER,
F.DEPARTMENT_NAME,G.BU_NAME
FROM USERS A INNER JOIN USER_PERSON B ON A.USER_ID=B.USER_ID
INNER JOIN TYPE_PERSON C ON C.TYPE_PERSON_ID = B.TYPE_PERSON_ID
INNER JOIN USER_TRANSACTION D ON D.USER_ID = A.USER_ID
INNER JOIN TRANSACTIONS E ON E.TRANSACTION_ID = D.TRANSACTION_ID
INNER JOIN DEPARTMENT F ON F.DEPARTMENT_ID = D.DEPARTMENT_ID
INNER JOIN BUS_UNIT G ON G.BU_ID = D.BU_ID
INNER JOIN BRANCH H ON H.BRANCH_ID = D.BRANCH_ID
INNER JOIN POSITION J ON J.POSITION_ID = D.POSITION_ID
WHERE A.FLAG = 'TRUE'
END
the result will select max(B.START_DATE) and max(D.START_DATE) and Group by USER_ID
try this
SELECT T.USER_ID ,
MAX(T.START_DATE) AS [Max First Start Date] ,
MAX(T.[Second Start Date]) AS [Max Second Start Date]
FROM ( SELECT A.USER_ID ,
B.START_DATE ,
D.START_DATE AS [Second Start Date] ,
A.FULL_NAME ,
A.COST_CENTER ,
F.DEPARTMENT_NAME ,
G.BU_NAME
FROM USERS A
INNER JOIN USER_PERSON B ON A.USER_ID = B.USER_ID
INNER JOIN TYPE_PERSON C ON C.TYPE_PERSON_ID = B.TYPE_PERSON_ID
INNER JOIN USER_TRANSACTION D ON D.USER_ID = A.USER_ID
INNER JOIN TRANSACTIONS E ON E.TRANSACTION_ID = D.TRANSACTION_ID
INNER JOIN DEPARTMENT F ON F.DEPARTMENT_ID = D.DEPARTMENT_ID
INNER JOIN BUS_UNIT G ON G.BU_ID = D.BU_ID
INNER JOIN BRANCH H ON H.BRANCH_ID = D.BRANCH_ID
INNER JOIN POSITION J ON J.POSITION_ID = D.POSITION_ID
WHERE A.FLAG = 'TRUE'
) AS T
GROUP BY T.USER_ID

mysql query - count words in common between users

I've a very nice query that selects friends of the current user. user_id = 2 in the example. His friend is user_id = 4.
I want the same query to fetch the number of words user_id 2 has with selected friends. In this case they have word = love, and this is also word they both have, so I want in_common row to say = 1.
Is it possible without changing too much current query?
Should I start from scratch?
SQL FIDDLE
Assuming that both users would have an entry for 'love' in the words_en table then something like this maybe:-
SELECT b.name_surname,
b.avatar,
b.friend_words,
(b.friend_msg_id) AS friend_msg_id,
words_common.words_in_common,
COUNT(m.id) AS unread_msg
FROM
(
SELECT a.name_surname as name_surname,
a.avatar as avatar,
GROUP_CONCAT(DISTINCT w.word ORDER BY w.word ASC) AS friend_words,
(a.friend_id) AS friend_msg_id
FROM
(
SELECT f1.asked_user_id AS friend_id,
f1.created,
u.name_surname,
u.avatar
FROM friends AS f1
INNER JOIN friends AS f2
ON f1.asked_user_id = f2.asker_user_id
AND f1.asker_user_id = f2.asked_user_id
INNER JOIN users AS u ON f1.asked_user_id = u.id
WHERE f1.status = 1 AND f2.status = 1
AND f1.asker_user_id = 2
) a
LEFT JOIN connections c ON c.user_id = a.friend_id
AND c.invisible <> 1 AND c.deleted <> 1
LEFT JOIN words_en w ON c.word_id = w.id
GROUP BY 1
) b
LEFT JOIN messages m ON m.to_user_id = 2
AND m.from_user_id = b.friend_msg_id
AND m.seen = 0
LEFT OUTER JOIN
(
SELECT b.user_id AS friend_id, GROUP_CONCAT(a.word) AS words_in_common
FROM words_en a
INNER JOIN words_en b
ON a.word = b.word
WHERE a.user_id = 2
GROUP BY b.user_id
) words_common
ON b.friend_msg_id = words_common.friend_id
GROUP BY b.name_surname, b.avatar, b.friend_words, b.friend_msg_id
ORDER BY unread_msg DESC
EDIT - modification to use connections table to find common words:-
SELECT b.name_surname,
b.avatar,
b.friend_words,
(b.friend_msg_id) AS friend_msg_id,
words_common.words_in_common,
COUNT(m.id) AS unread_msg
FROM
(
SELECT a.name_surname as name_surname,
a.avatar as avatar,
GROUP_CONCAT(DISTINCT w.word ORDER BY w.word ASC) AS friend_words,
(a.friend_id) AS friend_msg_id
FROM
(
SELECT f1.asked_user_id AS friend_id,
f1.created,
u.name_surname,
u.avatar
FROM friends AS f1
INNER JOIN friends AS f2
ON f1.asked_user_id = f2.asker_user_id
AND f1.asker_user_id = f2.asked_user_id
INNER JOIN users AS u ON f1.asked_user_id = u.id
WHERE f1.status = 1 AND f2.status = 1
AND f1.asker_user_id = 2
) a
LEFT JOIN connections c ON c.user_id = a.friend_id
AND c.invisible <> 1 AND c.deleted <> 1
LEFT JOIN words_en w ON c.word_id = w.id
GROUP BY 1
) b
LEFT JOIN messages m ON m.to_user_id = 2
AND m.from_user_id = b.friend_msg_id
AND m.seen = 0
LEFT OUTER JOIN
(
SELECT b.user_id AS friend_id, GROUP_CONCAT(c.word) AS words_in_common
FROM connections a
INNER JOIN connections b
ON a.word_id = b.word_id
INNER JOIN words_en c
ON b.word_id = c.id
WHERE a.user_id = 2
GROUP BY b.user_id
) words_common
ON b.friend_msg_id = words_common.friend_id
GROUP BY b.name_surname, b.avatar, b.friend_words, b.friend_msg_id
ORDER BY unread_msg DESC

Query goes wrong when checking if the user exists in other table

I have a query that shows a list of revisions and employees for every revision..
Now I'm trying to show if the given employee already has a row in the answers table..
This is the overview of the database
This is my working query that shows the list of revisions and employees
SELECT l.id, l.naam, r.id as revision_id, r.beschrijving, e.id as employee_id, e.voornaam, e.achternaam,
FROM lists l
INNER JOIN revisions r ON l.id = r.list_id
INNER JOIN employeelists el ON el.list_id= l.id
INNER JOIN employees e ON e.id = el.employee_id
INNER JOIN customers c ON c.id = e.customer_id
WHERE customer_id = :id AND r.actief = 1
Now I've tried several things to see if the employee already has a record in the answers table.. But It's failing the whole time..
Try 1 : Adding the Answers table with a left outer join
SELECT l.id, l.naam, r.id as revision_id, r.beschrijving, e.id as employee_id, e.voornaam, e.achternaam,
**CASE WHEN a.coach_id != 0 THEN 1 ELSE 0 END as FILLED IN**
FROM lists l"""
INNER JOIN revisions r ON l.id = r.list_id
**LEFT OUTER JOIN answers a ON a.revision_id = r.id**
INNER JOIN employeelists el ON el.list_id= l.id
INNER JOIN employees e ON e.id = el.employee_id
INNER JOIN customers c ON c.id = e.customer_id
WHERE customer_id = :id AND r.actief = 1
now the problem is that every employee is shown multiple times...
This is the SQLFiddle of the working database, The only thing i can't do is check if the given employee ( werknemer ) exists in the answers ( antwoorden ) table..
http://sqlfiddle.com/#!2/0c01c/4
Any idea on how i can solve this? I tried a subquery, but that didn't work out either.. Thanks!
Problem with Query Now
I thought i was correct but there's one more error. in the answers table, it shows results for werknemer_id ( employee_id ) = 78. For the revisie ( revision ) 1 and 2
While there is only results for revisie 1 (screenshot below)
Thanks!
How about this exist column will have 0 if not in antwoorden and 1 if exist
SELECT l.id, l.naam, r.revisie as revisie, r.id as revisie_id, r.beschrijving, w.id as werknemer, w.voornaam, w.achternaam
, a.werknemer_id,
(CASE WHEN a.werknemer_id IS NULL THEN 0 ELSE 1 END ) AS `exist`
FROM lijsten l
INNER JOIN revisies r ON l.id = r.lijst_id
INNER JOIN werknemerlijsten wl ON wl.lijst_id = l.id
INNER JOIN werknemers w ON w.id = wl.werknemer_id
INNER JOIN klanten k ON k.id = w.klant_id
LEFT JOIN antwoorden a ON w.id = a.werknemer_id
WHERE klant_id = 39 AND r.actief = 1
GROUP BY r.beschrijving, w.id
Fiddle
Final solution
This is the final outcome, only see 'ingevuld' is 1 if the desired coach (1 in thise case) is in the answers table.
SELECT l.id, l.naam, r.revisie AS revisie, r.id AS revisie_id, r.beschrijving, w.id AS werknemer, w.voornaam, w.achternaam, a.coach_id,
CASE WHEN a.coach_id = 1 THEN 1 ELSE 0 END AS ingevuld
FROM lijsten l
INNER JOIN revisies r ON l.id = r.lijst_id
INNER JOIN werknemerlijsten wl ON wl.lijst_id = l.id
INNER JOIN werknemers w ON w.id = wl.werknemer_id
INNER JOIN klanten k ON k.id = w.klant_id
LEFT JOIN antwoorden a ON w.id = a.werknemer_id AND r.id=a.revisie_id
WHERE klant_id = 39 AND r.actief = 1
group by r.id, w.id, a.coach_id