I have two main tables that comprise bookings for events.
A Registrants table (Bookings) R and an Events table E.
There are also two connected tables, Field_Values V and Event_Categories C
This diagram shows the relationship
What I am trying to do is create an Invoice query that mirrors the user's shopping cart. Often a user will book multiple events in one transaction, so my invoice should have columns for the common items e.g. User Name, User Email, Booking Date, Transaction ID and aggregated columns for the invoice line item values e.g. Quantity "1,2" Description "Desc1, Desc2" Price "10.00, 20.00" where there are two line items in the shopping cart.
The Transaction ID (dcea4_eb_registrant.transaction_id) is unique per Invoice and repeated per line item in that sale.
I have the following query which produces rows for each line item
SELECT
R.id as ID,
E.event_date as ServiceDate,
E.event_date - INTERVAL 1 DAY as DueDate,
Concat('Ad-Hoc Booking:',E.title) as ItemProductService,
Concat(R.first_name, ' ',R.last_name) as Customer,
R.first_name as FirstName,
R.last_name as LastName,
R.email,
R.register_date as InvoiceDate,
R.amount as ItemAmount,
R.comment,
R.number_registrants as ItemQuantity,
R.transaction_id as InvoiceNo,
R.published as Status,
E.event_date AS SERVICEDATE,
Concat('Ad-Hoc Booking:',E.title) AS DESCRIPTION,
R.number_registrants AS QUANTITY,
FORMAT(R.amount / R.number_registrants,2) AS RATE,
R.amount AS AMOUNT,
C.category_id as CLASS,
Concat(Group_Concat(V.field_value SEPARATOR ', '),'. ',R.comment) as Memo
FROM dcea4_eb_events E
LEFT JOIN dcea4_eb_registrants R ON R.event_id = E.id
LEFT JOIN dcea4_eb_field_values V ON V.registrant_id = R.id
LEFT JOIN dcea4_eb_event_categories C ON C.event_id = R.event_id
WHERE 1=1
AND V.field_id IN(14,26,27,15)
AND R.published <> 2 /*Including this line omits Cancelled Invoices */
AND R.published IS NOT NULL
AND (R.published = 1 OR R.payment_method = "os_offline")
AND (R.register_date >= CURDATE() - INTERVAL 14 DAY)
GROUP BY E.event_date, E.title, R.id, R.first_name, R.last_name, R.email,R.register_date, R.amount, R.comment
ORDER BY R.register_date DESC, R.transaction_id
This produces output like this
I'm using the following query to try to group together the rows with a common transaction_ID (rows two and three in the last picture) - I add group_concat on the columns I want to aggregate and change the Group By to be the transaction_id
SELECT
R.id as ID,
E.event_date as ServiceDate,
E.event_date - INTERVAL 1 DAY as DueDate,
Concat('Ad-Hoc Booking:',E.title) as ItemProductService,
Concat(R.first_name, ' ',R.last_name) as Customer,
R.first_name as FirstName,
R.last_name as LastName,
R.email,
R.register_date as InvoiceDate,
R.amount as ItemAmount,
R.comment,
R.number_registrants as ItemQuantity,
R.transaction_id as InvoiceNo,
R.published as Status,
Group_ConCat( E.event_date) AS SERVICEDATE,
Group_ConCat( Concat('Ad-Hoc Booking:',E.title)) AS DESCRIPTION,
Group_ConCat( R.number_registrants) AS QUANTITY,
Group_ConCat( FORMAT(R.amount / R.number_registrants,2)) AS RATE2,
Group_ConCat( R.amount) AS AMOUNT,
Group_ConCat( C.category_id) as CLASS,
Concat(Group_Concat(V.field_value SEPARATOR ', '),'. ',R.comment) as Memo
FROM dcea4_eb_events E
LEFT JOIN dcea4_eb_registrants R ON R.event_id = E.id
LEFT JOIN dcea4_eb_field_values V ON V.registrant_id = R.id
LEFT JOIN dcea4_eb_event_categories C ON C.event_id = R.event_id
WHERE 1=1
AND V.field_id IN(14,26,27,15)
AND R.published <> 2 /*Including this line omits Cancelled Invoices */
AND R.published IS NOT NULL
AND (R.published = 1 OR R.payment_method = "os_offline")
AND (R.register_date >= CURDATE() - INTERVAL 14 DAY)
GROUP BY R.transaction_id
ORDER BY R.register_date DESC, R.transaction_id
But this produces this output
It seems to be multiplying the rows. The Quantity column in the first row should just be 1 and in the second row it should be 2,1 .
I've tried using Group_Concat with DISTINCT but this doesn't work because often the values being concatenated are the same (e.g. the price for two events being booked are both the same) and the query only returns one value e.g. 10 and not 10, 10. The latter being what I need.
I'm guessing the issue is around the way the tables are joined but I'm struggling to work out how to get what I need.
Pointers in the right direction most appreciated.
You seem determined to go in what seems to me to be the wrong direction, so here's a gentle nudge down that hill...
Consider the following...
CREATE TABLE users
(user_id SERIAL PRIMARY KEY
,username VARCHAR(12) UNIQUE
);
INSERT INTO users VALUES
(101,'John'),(102,'Paul'),(103,'George'),(104,'Ringo');
DROP TABLE IF EXISTS sales;
CREATE TABLE sales
(sale_id SERIAL PRIMARY KEY
,purchaser_id INT NOT NULL
,item_code CHAR(1) NOT NULL
,quantity INT NOT NULL
);
INSERT INTO sales VALUES
( 1,101,'A',1),
( 2,103,'A',2),
( 3,103,'A',3),
( 4,104,'A',1),
( 5,104,'A',2),
( 6,104,'A',3),
( 7,103,'B',2),
( 8,103,'B',2),
( 9,104,'B',3),
(10,103,'B',2),
(11,104,'B',2),
(12,104,'B',1);
SELECT u.*
, x.sale_ids
, x.item_codes
, x.quantities
FROM users u
LEFT
JOIN
( SELECT purchaser_id
, GROUP_CONCAT(sale_id ORDER BY sale_id) sale_ids
, GROUP_CONCAT(item_code ORDER BY sale_id) item_codes
, GROUP_CONCAT(quantity ORDER BY sale_id) quantities
FROM sales
GROUP
BY purchaser_id
) x
ON x.purchaser_id = u.user_id;
+---------+----------+---------------+-------------+-------------+
| user_id | username | sale_ids | item_codes | quantities |
+---------+----------+---------------+-------------+-------------+
| 101 | John | 1 | A | 1 |
| 102 | Paul | NULL | NULL | NULL |
| 103 | George | 2,3,7,8,10 | A,A,B,B,B | 2,3,2,2,2 |
| 104 | Ringo | 4,5,6,9,11,12 | A,A,A,B,B,B | 1,2,3,3,2,1 |
+---------+----------+---------------+-------------+-------------+
I'm currently new to queries and below I have a query that I have made
QUERY:
select TITLE_ID, TITLE,NAME, JOB_CATEGORY AS ROLE
FROM MOVIES
NATURAL JOIN NEW_NAMES
WHERE JOB_CATEGORY = 'writer'
OR JOB_CATEGORY = 'director'
ORDER BY TITLE_ID ASC;
Which Displays:
TITLE_ID | TITLE | NAME | ROLE |
753595 | 2F2F | ROB | WRITER |
753595 | 2F2F | YAS | DIRECTOR|
However I would like it to display in this format below:
TITLE_ID | TITLE | WRITER | DIRECTOR|
753595 | 2F2F | ROB | YAS |
You join your NEW_NAMES table in twice. Once for director, once for writer. When you join a table more than once it's necessary that you give the table an Alias, here we use writer and director.
select TITLE_ID, TITLE,writer.NAME as writer_name, director.NAME as director_name
FROM MOVIES
LEFT OUTER JOIN NEW_NAMES as writer
ON MOVIES.TITLE_ID = writer.TITLE_ID
AND writer.JOB_CATEGORY = 'writer'
LEFT OUTER JOIN NEW_NAMES as director
ON MOVIES.TITLE_ID = director.TITLE_ID
AND director.JOB_CATEGORY = 'director'
ORDER BY TITLE_ID ASC;
I've made some assumptions in those ON clauses about which columns in your tables you are joining on. You may need to edit that.
According to error message you got, it is about Oracle, not MySQL.
Anyway, such a (classic?) principle should work: aggregate!
SQL> with
2 -- sample data
3 movies (title_id, title_name) as
4 (select 753595, '2F2F' from dual union all
5 select 123456, '1Z1Z' from dual
6 ),
7 new_names (title_id, job_category, name) as
8 (select 753595, 'writer', 'ROB' from dual union all
9 select 753595, 'director', 'YAS' from dual
10 )
11 -- query you need
12 select m.title_id,
13 m.title_name,
14 max(case when n.job_category = 'writer' then n.name end) as writer,
15 max(case when n.job_category = 'director' then n.name end) as director
16 from movies m left join new_names n on m.title_id = n.title_id
17 group by m.title_id, m.title_name
18 order by m.title_id;
TITLE_ID TITL WRI DIR
---------- ---- --- ---
123456 1Z1Z
753595 2F2F ROB YAS
SQL>
I want to create something like Facebook newsfeed: a person who follows other persons see their last posts at the top. I use this query as db select query but it has some problem.
Does anyone have any idea to improve this query?
These are important tables for this query:
______
| wall |
-------------------------------------------------------------------------------------
postID | inserterID | sportID | catID | title|insertDate|context|abstraction|editeDate
-------------------------------------------------------------------------------------
|follower|
---------------------
ID | foerID | foingID
---------------------
SELECT wall.postID
, wall.inserterID
, wall.title
, wall.insertDate
, wall.editDate
, wall.context
, user.lastName
, cat.catName
, wall.catID
, pic.picname AS pic
, foerID
FROM wall
JOIN user
ON user.userID = wall.inserterID
JOIN postcategory cat
ON cat.catID = wall.catID
LEFT
JOIN wallpic pic
ON pic.postID = wall.postID
LEFT
JOIN follow
ON follow.foingID = wall.inserterID
WHERE wall.inserterID != 71
GROUP
BY wall.postID
ORDER
BY follow.foerID IN (71) DESC
, wall.insertDate DESC
LIMIT 0,100
I've got three tables essentially:
hunt_c_users
hunt_c_collected_eggs
hunt_c_achievements
I'm grabbing a leaderboard currently only using hunt_c_users with a LEFT JOIN on hunt_c_collected_eggs but now I need to include hunt_c_achievements points also.'
Currently my query is:
SELECT
hunt_c_users.id AS 'id',
CONCAT(first_name, ' ', LEFT(last_name,1), '. From ', city) AS 'user_byline',
fbid AS 'user_fbid',
SUM(hunt_c_collected_eggs.value) AS user_points
FROM hunt_c_users
LEFT JOIN hunt_c_collected_eggs ON hunt_c_users.id = hunt_c_collected_eggs.user_id
WHERE hunt_c_users.id NOT IN ($bannedIDSstr)
GROUP BY hunt_c_users.id
HAVING SUM(hunt_c_collected_eggs.value) > 0
ORDER BY user_points DESC
LIMIT 10
But when I add in another LEFT JOIN on hunt_c_achievements and another SUM(hunt_c_achievements.value) the user_points is wildly inaccurate, and I understand why that's happening (the tables are joining horizontally so the achievements are added on each item), however I don't quite understand how to fix it.
This is the query that is giving me inaccurate results:
SELECT
hunt_c_users.id AS 'id',
CONCAT(first_name, ' ', LEFT(last_name,1), '. From ', city) AS 'user_byline',
fbid AS 'user_fbid',
SUM(hunt_c_collected_eggs.value) + SUM(hunt_c_achievements.value) AS user_points
FROM hunt_c_users
LEFT JOIN hunt_c_collected_eggs ON hunt_c_users.id = hunt_c_collected_eggs.user_id
LEFT JOIN hunt_c_achievements ON hunt_c_users.id = hunt_c_achievements.user_id
WHERE hunt_c_users.id NOT IN ($bannedIDSstr)
GROUP BY hunt_c_users.id
HAVING SUM(hunt_c_collected_eggs.value) > 0
ORDER BY user_points DESC
LIMIT 10
UPDATE
I have added a SQL Fiddle here: http://sqlfiddle.com/#!2/ac1f69/1
My expected result using the dummy-data provided is:
id | user_byline | user_fbid | user_points
------------------------------------------
1033 | ... | ... | 111
1030 | ... | ... | 97
1031 | ... | ... | 62
1032 | ... | ... | 27
You have to use subqueries to calculate the sums separately:
SET #bannedIDSstr = 0;
SELECT
user.id AS 'id',
CONCAT(user.first_name, ' ', LEFT(user.last_name, 1), '. From ', user.city) AS 'user_byline',
user.fbid AS 'user_fbid',
IFNULL(eggs_sum.collected_eggs, 0) + IFNULL(achievement_sum.achievement, 0) AS user_points
FROM hunt_c_users AS user
LEFT JOIN
(SELECT
eggs.user_id,
SUM(eggs.value) AS collected_eggs
FROM hunt_c_collected_eggs AS eggs
GROUP BY eggs.user_id
HAVING collected_eggs > 0
) AS eggs_sum
ON user.id = eggs_sum.user_id
LEFT JOIN
(SELECT
achievements.user_id,
SUM(achievements.value) AS achievement
FROM hunt_c_achievements AS achievements
GROUP BY achievements.user_id
) AS achievement_sum
ON user.id = achievement_sum.user_id
WHERE user.id NOT IN (#bannedIDSstr)
GROUP BY user.id
ORDER BY user_points DESC
LIMIT 10;
DEMO # SQL Fiddle
Thank you in advance for anyone who helps with this. I know i've seen this done before without too much pain but cant seem to find the solution
My database looks something like:
`tbl_user:
----------
id ( pkey )
email
fName
lName
tbl_userSparseType:
-------------------
id ( pkey )
varName
displayName
tbl_userSparse:
---------------
id ( pkey )
value ( Value of Sparse Type )
user_id ( => tbl_user.id )
userSparseType_id ( => tbl_userSparseType.id )
with sample data:
tbl_user:
(id, email, fName, lName)
1 Bob#example.com Billy Bob
2 Betty#example.com Betty Sue
3 Beam#example.com Jim Beam
tbl_userSparseType:
(id, varName, displayName)
1 fullName Full Name
2 dayPhone Day Phone
3 nightPhone Night Phone
4 cellPhone Cell Phone
5 homeAddr Home Address
tbl_userSparse:
(id, value, user_id, userSparseType_id)
1 Billy John Bob 1 1
2 James B. Beam 3 1
3 123-234-3456 1 2
4 234-345-4567 1 4
5 24 Best st. 2 5
6 456-567-6789 3 3
I tried doing two left joins, but this gave me a tbl_user row for each sparse entry like:
(id, email, fName, lName, displayName, value)
1,"Bob#example.com","Billy","Bob","Full Name","Billy John Bob"
1,"Bob#example.com","Billy","Bob","Day Phone","123-234-3456"
1,"Bob#example.com","Billy","Bob","Cell Phone","234-345-4567"
And despite a few 45 or so minute sessions of looking, I cant find a way to get something more like the following without explicitly naming the columns, I need a dynamic way to only pull all display names that apply to the subset of tbl_user rows being queried:
WHERE tbl_user.id IN (1,2)
id | email | fName | lName | Full Name, | Day Phone | Cell Phone |
Home Address
-------------------------------------------------------------------------------------------------------
1 | Bob#example.com | Billy | Bob | Billy John Bob | 123-234-3456 | 234-345-4567 |
2 | Betty#example.com | Betty | Sue | | | | 24 Best St.
Thanks again in advance, I'm hoping this can be done without too much fuss. :\
Unfortunately MySQL does not have a PIVOT function which is basically what you are trying to do. So you will need to use an aggregate function with a CASE statement. If you know the number of columns, then you can hard-code the values:
select u.id,
u.email,
u.fname,
u.lname,
max(case when t.displayname = 'Full Name' then us.value end) FullName,
max(case when t.displayname = 'Day Phone' then us.value end) DayPhone,
max(case when t.displayname = 'Cell Phone' then us.value end) CellPhone,
max(case when t.displayname = 'Home Address' then us.value end) HOmeAddress
from tbl_user u
left join tbl_userSparse us
on u.id = us.user_id
left join tbl_userSparseType t
on us.userSparseType_id = t.id
where u.id in (1, 2)
group by u.id, u.email, u.fname,u.lname;
See SQL Fiddle With Demo
Now if you want to perform this dynamically, meaning you do not know ahead of time the columns to transpose, then you should review the following article:
Dynamic pivot tables (transform rows to columns)
Your code would look like this:
SET #sql = NULL;
SELECT
GROUP_CONCAT(DISTINCT
CONCAT(
'max(case when t.displayname = ''',
t.displayname,
''' then us.value end) AS ',
replace(t.displayname, ' ', '')
)
) INTO #sql
FROM tbl_userSparse us
left join tbl_userSparseType t
on us.userSparseType_id = t.id;
SET #sql = CONCAT('SELECT u.id, u.email, u.fname, u.lname, ', #sql, '
from tbl_user u
left join tbl_userSparse us
on u.id = us.user_id
left join tbl_userSparseType t
on us.userSparseType_id = t.id
where u.id in (1, 2)
group by u.id, u.email, u.fname, u.lname');
PREPARE stmt FROM #sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;
See SQL Fiddle with Demo