GROUP_CONCAT With Nested Set Model - php

I have an application that uses a nested set model class to organise my data, however I'm trying to write a query that will group_concat my results. I know I need to put some sub select statements somewhere but I can't figure it out!
Here's my structure at the moment:
table: person
-----------+------------+-----------
|Person_ID | Name | Age |
-----------+------------+-----------
| 1 | Mark Vance | 19 |
| 2 | Michael Tsu| 22 |
| 3 | Mark Jones | 29 |
| 4 | Sara Young | 25 |
-----------+------------+-----------
table: person_to_group
----+------------+-----------
|ID | Person_ID | Group_ID |
----+------------+-----------
| 1 | 3 | 1 |
| 2 | 3 | 2 |
| 3 | 1 | 2 |
| 4 | 4 | 3 |
----+------------+-----------
table: groups
----------+--------------+--------------+-------------
|Group_ID | Group_Name | Group_Left | Group_Right |
----------+--------------+--------------+-------------
| 1 | Root | 1 | 6 |
| 2 | Node | 2 | 5 |
| 3 | Sub Node | 3 | 4 |
----------+--------------+--------------+-------------
I need to render something like this with my results:
//Grab the group_IDs for this person and put them in the class tag...
<li class="2 3">Sara Young is in the Sub Node Group</li>
Notice that although Sara is in the Sub Node group, she is still being given the id for Node aswell because she is a child of Node.
The following is the query that I am working with as a starting point.
SELECT *, GROUP_CONCAT( CAST( gg.Group_ID AS CHAR ) SEPARATOR ' ' ) Group_IDs
FROM groups gg
LEFT JOIN person_to_group AS t1 ON gg.Group_ID = t1.Group_ID
LEFT JOIN person AS t2 ON t2.Person_ID = t1.Person_ID
GROUP BY t2.per_ID
ORDER BY t2.Name ASC
Any help would be much appreciated!

Here's how I'd write the query:
SELECT p.Name,
GROUP_CONCAT( g.Group_Name ) AS Group_List,
GROUP_CONCAT( CAST( gg.Group_ID AS CHAR ) SEPARATOR ' ' ) AS Group_ID_List
FROM person AS p
INNER JOIN person_to_group AS pg ON p.Person_ID = pg.Person_ID
INNER JOIN groups AS g ON pg.Group_ID = g.Group_ID
INNER JOIN groups AS gg ON g.Group_Left BETWEEN gg.Group_Left AND gg.Group_Right
GROUP BY p.Name
ORDER BY p.Name ASC
Note that if you group by person name, you also need to GROUP_CONCAT the list of group names. According to your schema, a person could belong to multiple groups, because of the many-to-many relationship.
I also recommend against using SELECT * in general. Just specify the columns you need.

This was little bit interesting as I do programming in both MsSQL and MySql. In SQL I have used function called STUFF. In MySQL you can use a function called INSERT. I tried out the below query in MsSQL. Don't have a MySQL handy to try out my query. If I have time I will post the MySQL version of the query.
DECLARE #person TABLE (Person_ID INT, Name VARCHAR(50), Age INT)
INSERT INTO #person VALUES
(1,'Mark Vance',19),
(2,'Michael Tsu',22),
(3,'Mark Jones',29),
(4,'Sara Young',25)
DECLARE #groups TABLE (Group_ID INT, Group_Name VARCHAR(50), Group_Left INT, Group_Right INT)
INSERT INTO #groups VALUES
(1,'Root',1,6),
(2,'Node',2,5),
(3,'Sub Node',3,4)
DECLARE #person_to_group TABLE (ID INT, Person_ID INT, Group_ID INT)
INSERT INTO #person_to_group VALUES
(1,3,1),
(2,3,2),
(3,1,1),
(4,4,1),
(4,1,1)
SELECT *,STUFF((SELECT ',' + CAST(g.Group_ID AS VARCHAR) FROM #groups g
JOIN #person_to_group pg ON g.Group_ID = pg.Group_ID AND pg.Person_ID = a.Person_ID FOR XML PATH('')) , 1, 1, '' ) FROM #person a
Function: INSERT(str,pos,len,newstr)
Documentation: http://dev.mysql.com/doc/refman/5.0/en/string-functions.html#function_insert

Related

PDO Query : Count related and repeated values then inner join

I work with PHP and PDO.
So I have 2 tables like,
Table 1
| id | name | age |
| 1 | John | 25 |
| 2 | Tom | 32 |
| 3 | James| 45 |
Table 2
| id | Comment | Link |
| 1 | some text | 3 |
| 2 | some text | 3 |
| 3 | some text | 1 |
So, Link column numbers represent id's in table1. For example Link = 3s in table 2 represent James in table 1. I need a query which brings all table1's data and also a number of repeated value for related Link column which comes from table2.
For example, the query should give me (let's choose James),
| id | name | age | Value |
| 3 | James | 45 | 2 |
value=2, because there are two 3s in link column which related to James
I tried somethings but got lots of errors.
I think you just need the GROUP BY
SELECT a.id,
a.name,
a.age,
count(*) as value
FROM table1 a
JOIN table2 b ON a.id = b.link
GROUP BY a.id, a.name, a.age
If you really want just one row then add WHERE
SELECT a.id,
a.name,
a.age,
count(*) as value
FROM table1 a
JOIN table2 b ON a.id = b.link
WHERE a.name = 'James'
GROUP BY a.id, a.name, a.age
or use subquery
SELECT a.id,
a.name,
a.age,
(SELECT count(*) FROM table2 b WHERE a.id = b.link) as value
FROM table1 a
WHERE a.name = 'James'

MySql - Avoid SET() using junction and foreign keys

So here is the problem :
I have a table "Members" with members and their attributes (name, birthday, mail, etc.)
These members may belong to groups (let's say there are 3 groups), from none to all of them. And these groups are referenced in a table ("Groups") so I can add/delete/modify them as I want.
SET() doesn't seem to be a solution, it isn't compatible with foreign keys / reference table.
So at first, I was thinking of doing a TINYINT() column, which I use like SET() : 111 (7) for all groups, 000 (0) for none, 001 (1) for the 1st group , 010 (2) for the 2nd, etc. But since the names are quite complex, it's confusing, and not much more compatible with foreign keys.
I read that I should do a 3rd table "Members-Groups" with memberID and groupID to join both of my two tables, but I don't clearly understand how it work.
What I understand is that I will have a table with IDs of members and groups like this :
+----------+---------+
| memberID | groupID |
+----------+---------+
| 1 | 1 |
| 1 | 2 |
| 2 | 1 |
| 2 | 3 |
| 3 | 2 |
+----------+---------+
and combined with junction I can retrieve what I want. Is it right ? Otherwise can someone explain me how i should do ?
I precise that I'd like to have as final result (after sql request + php script) a member, his attributes and the groups he belongs to in a single row (as with SET()), even members that doesn't belong to any group.
Assuming
drop table if exists mg;
drop table if exists m;
create table m (id int primary key, name varchar(3));
insert into m values
(1,'abc'),
(2,'def'),
(3,'ghi');
drop table if exists g;
create table g(id int primary key ,name varchar(3));
insert into g values
(1,'aaa'),
(2,'bbb'),
(3,'ccc');
create table mg
(memid int,grid int,
index fmid(memid,grid) ,
foreign key (memid) references m(id) on delete cascade,
foreign key (grid) references g(id) on delete cascade
);
insert into mg values
(1,1),(1,2),(1,3),
(2,1),(2,3);
You could join the 3 tables and produce the results using group_concat or conditional aggregation.
MariaDB [sandbox]> select m.id,m.name, group_concat(g.name) groups
-> from m
-> join mg on mg.memid = m.id
-> join g on mg.grid = g.id
-> group by m.id,m.name;
+----+------+-------------+
| id | name | groups |
+----+------+-------------+
| 1 | abc | aaa,bbb,ccc |
| 2 | def | aaa,ccc |
+----+------+-------------+
2 rows in set (0.00 sec)
MariaDB [sandbox]>
MariaDB [sandbox]> select m.id,m.name,
-> max(case when g.id = 1 then g.name else '' end) as group1,
-> max(case when g.id = 2 then g.name else '' end) as group2,
-> max(case when g.id = 3 then g.name else '' end) as group3
-> from m
-> join mg on mg.memid = m.id
-> join g on mg.grid = g.id
-> group by m.id,m.name;
+----+------+--------+--------+--------+
| id | name | group1 | group2 | group3 |
+----+------+--------+--------+--------+
| 1 | abc | aaa | bbb | ccc |
| 2 | def | aaa | | ccc |
+----+------+--------+--------+--------+
2 rows in set (0.00 sec)
If you want members who don't belong to any group change the joins to left joins.
ariaDB [sandbox]> select m.id,m.name, group_concat(g.name) groups
-> from m
-> left join mg on mg.memid = m.id
-> left join g on mg.grid = g.id
-> group by m.id,m.name;
+----+------+-------------+
| id | name | groups |
+----+------+-------------+
| 1 | abc | aaa,bbb,ccc |
| 2 | def | aaa,ccc |
| 3 | ghi | NULL |
+----+------+-------------+
3 rows in set (0.00 sec)
MariaDB [sandbox]>
MariaDB [sandbox]> select m.id,m.name,
-> max(case when g.id = 1 then g.name else '' end) as group1,
-> max(case when g.id = 2 then g.name else '' end) as group2,
-> max(case when g.id = 3 then g.name else '' end) as group3
-> from m
-> left join mg on mg.memid = m.id
-> left join g on mg.grid = g.id
-> group by m.id,m.name;
+----+------+--------+--------+--------+
| id | name | group1 | group2 | group3 |
+----+------+--------+--------+--------+
| 1 | abc | aaa | bbb | ccc |
| 2 | def | aaa | | ccc |
| 3 | ghi | | | |
+----+------+--------+--------+--------+
3 rows in set (0.00 sec)
I feel half-confused by the question, but I'll take a stab at it.
If you have a Members table, then it makes sense to have member_id be the unique primary key. If you want to store which groups each member is in, simply add a new column to the Members table for each Group.
As for the values to be given to columns Group1, Group2, Group3 you could set them as ENUM('0','1') or ENUM('No','Yes') or whatever and make the default value the negative-meaning (first) value.
With this db structure, you won't have to bother chopping up a string during querying -- you just write SELECT or WHERE statements that specify the appropriate Group column value.
If this doesn't directly answer, please clarify your question.

MySQL - inner join - add column with value based on other value

I'm struggling with mysql joins :/
I've multiple tables inside database fe. tasks, users etc.
Table tasks containing tasks with various variables, but the most important - id's of users signed to task (as different roles inside the task - author, graphic, corrector):
+---------+-------------+--------------+
| task_id | task_author | task_graphic |
+---------+-------------+--------------+
| 444 | 1 | 2 |
+---------+-------------+--------------+
Table users
+---------+----------------+------------+-----------+
| user_id | user_nice_name | user_login | user_role |
+---------+----------------+------------+-----------+
| 1 | Nice Name #1 | login1 | 0 |
+---------+----------------+------------+-----------+
| 2 | Bad Name #2 | login2 | 1 |
+---------+----------------+------------+-----------+
Using PDO I'm getting the whole data I want while using INNER JOIN with data from different tables (and $_GET variable)
SELECT tasks.*, types.types_name, warehouse.warehouse_id, warehouse.warehouse_code, warehouse.warehouse_description
FROM tasks
INNER JOIN types ON types.types_id = tasks.task_id
INNER JOIN warehouse ON warehouse.warehouse_id = tasks.task_id
WHERE tasks.task_id = '".$get_id."'
ORDER BY tasks.task_id
Above query returns:
+---------+--------------+--------------+----------------+------------+-----------+------------------+------------------------+------------+-------------+-----------------+-----------+----------------+--------------------+---------------------+-----------+---------------------+------------------+---------------------+
| task_id | task_creator | task_graphic | task_purchaser | task_title | task_lang | task_description | task_description_files | task_files | task_status | task_prod_index | task_type | task_print_run | task_print_company | task_warehouse_code | task_cost | task_time_added | task_deadline | task_date_warehouse |
+---------+--------------+--------------+----------------+------------+-----------+------------------+------------------------+------------+-------------+-----------------+-----------+----------------+--------------------+---------------------+-----------+---------------------+------------------+---------------------+
| 2 | 1 | 2 | 1 | Test | PL | Lorem ipsum (?) | | | w | 2222 | 3 | 456546 | Firma XYZ | 2 | 124 | 29.09.2016 15:48:20 | 01.10.2016 12:00 | 07.10.2016 14:00 |
+---------+--------------+--------------+----------------+------------+-----------+------------------+------------------------+------------+-------------+-----------------+-----------+----------------+--------------------+---------------------+-----------+---------------------+------------------+---------------------+
And I'd like to get query with added user_nice_name after task_creator, task_author and task_graphic - obviously nice names selected from table users based on ID's provide in 3 above fields fe.
+---------+--------------+------------------------------------+--------------+--------------------------------------+
| task_id | task_creator | task_creator_nn | task_graphic | task_graphic |
+---------+--------------+------------------------------------+--------------+--------------------------------------+
| 2 | 1 | Nice Name (from task_creator ID=1) | 2 | Nice Name (from task_graphic ID = 2) |
+---------+--------------+------------------------------------+--------------+--------------------------------------+
How can I achieve that?
You need three joins:
SELECT t.*,
uc.user_nice_name as creator_name,
ug.user_nice_name as graphic_name,
up.user_nice_name as purchaser_name,
ty.types_name, w.warehouse_id, w.warehouse_code, w.warehouse_description
FROM tasks t INNER JOIN
types ty
ON ty.types_id = t.task_id INNER JOIN
warehouse w
ON w.warehouse_id = t.task_id LEFT JOIN
users uc
ON uc.user_id = t.task_creator LEFT JOIN
users ug
ON ug.user_id = t.task_graphic LEFT JOIN
users up
ON up.user_id = t.task_purchaser
WHERE t.task_id = '".$get_id."'
ORDER BY t.task_id;
Notes:
Table aliases make the query easier to write and to read. They are also required because you have three references to users in the FROM clause.
This uses LEFT JOIN for the users in case some of the reference values are missing.
You need to work on your naming. It doesn't make sense that a "warehouse" id matches a "task" id. Or that a "task" id matches a "types" id. But that is how you phrased the query in your question.
The ORDER BY effectively does nothing, because all rows have the same task_id.
Assuming that the task_graphic_name is inside a table name task_graphic_table and the relation field are task_graphic_id
SELECT tasks.*
, types.types_name
, warehouse.warehouse_id
, warehouse.warehouse_code
, warehouse.warehouse_description
, users.user_nice_name
FROM tasks
INNER JOIN types ON types.types_id = tasks.task_id
INNER JOIN warehouse ON warehouse.warehouse_id = tasks.task_id
INNER JOIN users ON users.user_nice_name = tasks.task_graphic
WHERE tasks.task_id = '".$get_id."'
ORDER BY tasks.task_id
And if you need the column appear in a specific order you should explicitally call the column name in sequence eg:
SELECT tasks.col1
, task.col2
, types.types_name
, warehouse.warehouse_id
, warehouse.warehouse_code
, task.col2
, warehouse.warehouse_description
, task_graphic_table.task_graphic_name
Add two sub query in with your query. like
SELECT tasks.*,
....
....,
(select user_nice_name from users where id = tasks.task_author) AS task_creator_name,
(select user_nice_name from users where id = tasks.task_graphic) AS task_graphic_name
FROM tasks
INNER JOIN types ON types.types_id = tasks.task_id
....
....

PHP MySQL joining two tables with conditional joins

So, I have a database that I am creating. It stores information about families and each family member. It then uses those records to associate invoices to either a family or family member.
My dilemma is that I need to list all of these invoices to a page under the families record i.e. create a list of invoices associated to either the family itself or an individual family member.
Table Structure
invoices
id | date_entered | invoice_date | invoice_number | invoice_amount | client_type | unique_id | supplier_type | supplier_id | category_id | childcare_hours
---+--------------+--------------+----------------+----------------+-------------+-----------+---------------+-------------+-------------+----------------
1 | 1411098397 | 1411048800 | 123 | 0.01 | 0 | 137 | 0 | 139 | 5 | NULL
families
id | ufi | last_name | address_1 | address_2 | city_id | phone | mobile | email | f_d_worker_1 | f_d_worker_2 | status_id | trans_date | entry_date | exit_date | eligible_date | active_date | lga_loc_id | facs_loc_id | ind_status_id | referral_id | active_status | comm_org_id | notes
---+----------+-----------------+-------------+-----------+---------+-------+--------+-------+--------------+--------------+-----------+------------+------------+-----------+---------------+-------------+------------+-------------+---------------+-------------+---------------+-------------+-------
1 | 1-XEWUDZ | Forsyth - Ennis | Skinner St. | NULL | NULL | NULL | NULL | NULL | 13 | NULL | 1 | NULL | 1341324000 | NULL | 1341842400 | 1342620000 | 7 | 1 | 3 | NULL | 1 | 1 | NULL
clients (family members)
id | upi | last_name | first_name | birthdate | sex | phone | mobile | email | indig_status_id | referral_id | relationship_id | preschool_id | family_id | notes
---+----------+-----------+------------+------------+-----+-------+--------+-------+-----------------+-------------+-----------------+--------------+-----------+------
1 | 1-XFCBBP | Ennis | Jason | 20/09/1996 | 1 | NULL | NULL | NULL | 3 | NULL | NULL | NULL | 1 | NULL
My current SQL looks like:
SELECT `invoices`.`id`, `invoices`.`date_entered`, `invoices`.`invoice_date`, `invoices`.`invoice_number`, `invoices`.`invoice_amount`, `invoices`.`client_type`, `invoices`.`unique_id`, `unique1`.`ufi`, `unique2`.`upi`, `unique1`.`last_name`, `invoices`.`supplier_type`, `invoices`.`supplier_id`, `suppliers`.`name`, `invoices`.`category_id`, `cat1`.`name`, `cat2`.`name`, `invoices`.`childcare_hours`
FROM `invoices`
LEFT OUTER JOIN `suppliers` ON `suppliers`.`id` = `invoices`.`supplier_id`
LEFT OUTER JOIN `categories` cat1 ON `cat1`.`id` = `invoices`.`category_id`
LEFT OUTER JOIN `preschool_types` cat2 ON `cat2`.`id` = `invoices`.`category_id`
LEFT OUTER JOIN `families` unique1 ON `unique1`.`id` = `invoices`.`unique_id`
LEFT OUTER JOIN `clients` unique2 ON `unique2`.`id` = `invoices`.`unique_id`
WHERE (`invoices`.`unique_id` = ? AND `unique1`.`ufi` = ?) LIMIT 0, 10
But what I need a query that checks the client_type column and if it equals 1 it needs to look in the clients table BUT it needs to look for members of the same family, identified by the id row in the families table
SOLUTION
Ok, so after much, much (much) screwing around and a little research. It appears that #cupid was correct (Although very brief in his answer).
And I will explain the solution better (in hope that this will help someone later).
The UNION option in MySQL (and most likely other SQL) allows you to combine the result sets of two (or more) SELECT queries, into one result set. This is extremely helpful if you have similar data, in separate tables that you may want to select easily and process as one request. Also helpful (in my case) for pagination, by allowing you to utilise SQL's LIMIT option.
One thing to take into consideration is that, the UNION syntax uses the columns from the first SELECT statement as the column names for all following queries, also you need to make sure that you have the same amount of columns selected in all queries for this to work.
(
SELECT
`invoices`.`id`,
`invoices`.`date_entered`,
`invoices`.`invoice_date`,
`invoices`.`invoice_number`,
`invoices`.`invoice_amount`,
`invoices`.`client_type`,
`invoices`.`unique_id`,
`clients`.`upi`,
`clients`.`last_name`,
`clients`.`family_id`,
`invoices`.`supplier_type`,
`invoices`.`supplier_id`,
`suppliers`.`name`,
`invoices`.`category_id`,
`cat1`.`name`,
`cat2`.`name`,
`invoices`.`childcare_hours`
FROM
(
`invoices`
LEFT OUTER JOIN `suppliers` ON `suppliers`.`id` = `invoices`.`supplier_id`
LEFT OUTER JOIN `categories` cat1 ON `cat1`.`id` = `invoices`.`category_id`
LEFT OUTER JOIN `preschool_types` cat2 ON `cat2`.`id` = `invoices`.`category_id`
LEFT OUTER JOIN `clients` ON `clients`.`id` = `invoices`.`unique_id`)
WHERE
`clients`.`family_id` = 47 AND `invoices`.`client_type` = 1
)
UNION
(
SELECT
`invoices`.`id`,
`invoices`.`date_entered`,
`invoices`.`invoice_date`,
`invoices`.`invoice_number`,
`invoices`.`invoice_amount`,
`invoices`.`client_type`,
`invoices`.`unique_id`,
`families`.`ufi`,
`families`.`last_name`,
`families`.`id`,
`invoices`.`supplier_type`,
`invoices`.`supplier_id`,
`suppliers`.`name`,
`invoices`.`category_id`,
`cat1`.`name`,
`cat2`.`name`,
`invoices`.`childcare_hours`
FROM `invoices`
LEFT OUTER JOIN `suppliers` ON `suppliers`.`id` = `invoices`.`supplier_id`
LEFT OUTER JOIN `categories` cat1 ON `cat1`.`id` = `invoices`.`category_id`
LEFT OUTER JOIN `preschool_types` cat2 ON `cat2`.`id` = `invoices`.`category_id`
LEFT OUTER JOIN `families` ON `families`.`id` = `invoices`.`unique_id`
WHERE
`invoices`.`unique_id` = 47 AND `invoices`.`client_type` = 0
)
Have you thought about using UNION?

JOIN rows from 3 tables WHERE tableA.column1 = tableB.column1 = tableC.column1

I have 3 tables, structured like so:
TABLE A:
ID | PCID | ACTIVE | COHORT | WEEKLY_MEETING_TIME | FYE_ID | RC | AGREEMENT_SIGNED | RELEASE_SIGNED | NOTES | FACULTY_ADVISOR
TABLE B:
ID | QUARTER | OFFICE | WRITING_CENTER | etc.. | etc.. | etc.. |
TABLE C:
ID | QUARTER | WEEK | EMAIL | etc.. | etc.. | etc.. |
The common element between all 3 tables is the ID field.
I need to SELECT from all 3 tables, and have each row represent one ID and all the values associated with that ID.
So, for example, each output row should look like a combination of the three tables:
RESULTS:
ID | PCID | ACTIVE | COHORT | WEEKLY_MEETING_TIME | FYE_ID | RC | AGREEMENT_SIGNED | RELEASE_SIGNED | NOTES | FACULTY_ADVISOR | QUARTER | OFFICE | WRITING_CENTER | etc.. | etc.. | etc.. | WEEK | EMAIL | etc.. | etc.. | etc.. |
I have no idea how to structure a query like this. I suspect it involves using JOINs but my attempts have proved futile.
how can I combine the data from 3 tables, based on shared ID field?
SELECT * FROM TABLEA
INNER JOIN TABLEB ON TABLEA.ID = TABLEB.ID
INNER JOIN TABLEC ON TABLEA.ID = TABLEC.ID
If you don't need all the values substitute the "*" with the names of the fields you need (ex. TABLEA.ID, TABLEB.QUARTER, TABLEC.WEEK...)
NATURAL JOIN works in MySQL:
select * from A natural join B natural join C
Try this :
SELECT / WHAT YOU NEED TO SELECT - NOT * !!!/
FROM TABLEA
INNER JOIN TABLEB USING (ID)
INNER JOIN TABLEC USING (ID)

Categories