Related
I know this question was already asked in a similar way, but I could not found any with an alias in the where clause.
I have a table structure like this:
CREATE TABLE Orders
( ID int NOT NULL Primary Key
, OrderNr VARCHAR(6) NOT NULL
, Date DATE NOT NULL
, Time CHAR(6) NOT NULL
, GeoCode CHAR(6) NULL) ;
My insert looks like this:
INSERT INTO orders (ID, OrderNr, Date, Time, GeoCode) VALUES (1, '123456', '2022-02-
15', '111110', '4022')
, (2, '123457', '2022-02-15', '121210', '4022')
, (3, '123455', '2021-04-15', '171515', '4020')
, (4, '123455', '2021-04-16', '150302', '4022')
, (5, '123466', '2022-03-03', '191810', '4020')
, (6, '123466', '2022-03-04', '121410', '4022')
Now I´m trying to get the latest Date and Time values for all OrderNr like this:
SELECT ID, OrderNr, MAX(cast(concat(Date, ' ', cast(Time as Time)) as datetime)) as
DateAndTime, GeoCode
FROM Orders o1
GROUP BY OrderNr
The Results shows the right latest date and time but the GeoCode is wrong. E.g for the
OrderNr 123455 it is 4020 but should be 4022.
I know that similar question were already asked but I cant use the alias in the where clause. Can somebody explain to me what I´m doing wrong?
Thank you very much in advance.
If your mysql version support ROW_NUMBER window function you can try this
SELECT *
FROM (
SELECT ID,
OrderNr,
cast(concat(Date, ' ', cast(Time as Time)) as datetime) DateAndTime,
GeoCode,
ROW_NUMBER() OVER(PARTITION BY OrderNr ORDER BY cast(concat(Date, ' ', cast(Time as Time)) as datetime) DESC) rn
FROM Orders o1
) t1
WHERE rn = 1
or use subquery with EXISTS
SELECT *
FROM Orders o1
WHERE EXISTS (
SELECT 1
FROM Orders oo
WHERE oo.OrderNr = o1.OrderNr
HAVING MAX(oo.Date) = o1.Date
)
sqlfiddle
This is my sample data
CREATE TABLE customer1
(
rating int(9),
genre varchar(100),
title varchar(100)
);
INSERT INTO customer1 (rating, genre, title)
VALUES
(2, 'A', 'abc'),
(4, 'A', 'abc1'),
(2, 'B', 'abc2'),
(3, 'B', 'abc3'),
(2, 'C', 'abc4'),
(5, 'C', 'abc5');
I need to find the title with max rating in each genre.
Thanks for the help.
One option uses a subquery for filtering:
select c.*
from customer1
where c.rating = (select max(c1.rating) from customer1 c1 where c1.genre = c.genre)
This would take advantage of an index on (genre, rating).
In MySQL 8.0, you can also use window functions:
select *
from (
select c.*,
rank() over(partition by genre order by rating desc) rn
from customer1 c
) c
where rn = 1
I have a column with complex user id. I want to replace the text within my select query.
This creates a new column as updated_by for every single value. I want them to be replaced in a single column. How can I achieve this?
select replace(updated_by, '5eaf5d368141560012161636', 'A'),
replace(updated_by, '5e79d03e9abae00012ffdbb3', 'B'),
replace(updated_by, '5e7b501e9abae00012ffdbd6', 'C'),
replace(updated_by, '5e7b5b199abae00012ffdbde', 'D'),
replace(updated_by, '5e7c817c9ca5540012ea6cba', 'E'),
updated_by
from my_table
GROUP BY updated_by;
In Postgres I would use a VALUES expression to form a derived table:
To just select:
SELECT *
FROM my_table m
JOIN (
VALUES
('5eaf5d368141560012161636', 'A')
, ('5e79d03e9abae00012ffdbb3', 'B')
, ('5e7b501e9abae00012ffdbd6', 'C')
, ('5e7b5b199abae00012ffdbde', 'D')
, ('5e7c817c9ca5540012ea6cba', 'E')
) u(updated_by, new_value) USING (updated_by);
Or LEFT JOIN to include rows without replacement.
You may need explicit type casts with non-default data types. See:
Casting NULL type when updating multiple rows
For repeated use, create a persisted translation table.
CREATE TABLE updated_by_translation (updated_by text PRIMARY KEY, new_value text);
INSERT INTO my_table
VALUES
('5eaf5d368141560012161636', 'A')
, ('5e79d03e9abae00012ffdbb3', 'B')
, ('5e7b501e9abae00012ffdbd6', 'C')
, ('5e7b5b199abae00012ffdbde', 'D')
, ('5e7c817c9ca5540012ea6cba', 'E')
;
Data types and constraints according to your actual use case.
SELECT *
FROM my_table m
LEFT JOIN updated_by_translation u USING (updated_by);
MySQL recently added a VALUES statement, too. The manual:
VALUES is a DML statement introduced in MySQL 8.0.19
But it requires the keyword ROW for every row. So:
...
VALUES
ROW('5eaf5d368141560012161636', 'A')
, ROW('5e79d03e9abae00012ffdbb3', 'B')
, ROW('5e7b501e9abae00012ffdbd6', 'C')
, ROW('5e7b5b199abae00012ffdbde', 'D')
, ROW('5e7c817c9ca5540012ea6cba', 'E')
...
Use case:
select case updated_by
when '5eaf5d368141560012161636' then 'A'
when '5e79d03e9abae00012ffdbb3' then 'B'
when '5e7b501e9abae00012ffdbd6' then 'C'
when '5e7b5b199abae00012ffdbde' then 'D'
when '5e7c817c9ca5540012ea6cba' then 'E'
end as updated_by
from my_table
This has to be nested liek this
SELECT
REPLACE(REPLACE(REPLACE(REPLACE(REPLACE(updated_by,
'5e7c817c9ca5540012ea6cba',
'E'),
'5e7b5b199abae00012ffdbde',
'D'),
'5e7b501e9abae00012ffdbd6',
'C'),
'5e79d03e9abae00012ffdbb3',
'B'),
'5eaf5d368141560012161636',
'A'),
updated_by
FROM
my_table
GROUP BY updated_by
This will replace all occurring, patterns, if they are not foung nothing happens
You can use a recursive CTE if you need to handle multiple values within a single row:
with replacements as (
select '5eaf5d368141560012161636' as oldval, 'A' as newval union all
select '5e79d03e9abae00012ffdbb3' as oldval, 'B' union all
select '5e7b501e9abae00012ffdbd6' as oldval, 'C' union all
select '5e7b5b199abae00012ffdbde' as oldval, 'D' union all
select '5e7c817c9ca5540012ea6cba' as oldval, 'E'
),
r as (
select r.*, row_number() over (order by oldval) as seqnum
from replacements r
),
recursive cte (
select r.seqnum, replace(t.updated_by, r.oldval, r.newval) as updated_by
from my_table t join
r
on r.seqnum = 1
union all
select r.seqnum, replace(cte.updated_by, r.oldval, r.newval) as updated_by
from cte t join
r
on r.seqnum = cte.seqnum + 1
)
select cte.*
from cte
where seqnum = (select count(*) from replacements);
I want to get the average number (Attendee NPS) from a SQL table I've already put together.
I've encased the initial table in a new select statement so I can take the average of distinct values. Is there something in my Join clause that is preventing this from working?
Im getting the following error:
ERROR: missing FROM-clause entry for table "gizmo" Position: 12
SELECT
avg(bigtable.gizmo.attendee_nps)
FROM
(
SELECT DISTINCT
attendee_survey_results.swoop_event_id AS "Swoop ID",
attendee_survey_results.startup_weekend_city AS "SW City",
swooptable.start_date AS "Date",
gizmo.attendee_nps AS "Attendee NPS"
FROM attendee_survey_results
JOIN
(
SELECT
swoop_event_id,
(
100 * count(CASE WHEN attendee_nps >= 9 THEN 1 END)
/ count(attendee_nps)
- 100 * count(CASE WHEN attendee_nps <= 6 THEN 1 END)
/ count(attendee_nps)
) AS "attendee_nps"
FROM attendee_survey_results
GROUP BY swoop_event_id
) AS "gizmo"
ON gizmo.swoop_event_id = attendee_survey_results.swoop_event_id
JOIN
(
SELECT eid,start_date,manager_email
FROM events
) AS "swooptable"
ON gizmo.swoop_event_id = swooptable.eid
) AS bigtable
[edit, ok you don't have a single problem, but the request at the bottom should work]
3 part notation bigtable.gizmo.attendee_nps
You can't use this bigtable.gizmo.attendee_nps, this is the "with DB" specific syntax : db_name.tbl_name.col_name.
You should use a table_or_alias.col_name_or_alias notation
In sub query you loose the deep table name of every deep-1 :
-- with the deep explicite
SELECT `d0`.`new_field`
FROM (
-- implicite `d1` table
SELECT `new_field`
FROM (
-- with the deep `d2` explicite and alias of field
SELECT `d2`.`field` AS `new_field`
FROM (
-- without the explicite `d3` table and `field` field
SELECT *
FROM (
-- output a `field` => 12
SELECT 12 as `field`
) AS `d3`
) AS `d2`
) AS `d1`
) AS `d0`
-- print `new_field` => 12
Access deep-1 aliased field
SELECT `attendee_nps`
FROM
(
SELECT `attendee_nps` AS `new_alias_field`
FROM attendee_survey_results
) AS bigtable
Unknown column 'attendee_nps' in 'field list'
When you make a field alias in deep-1 query, deep-0 can only access the alias new_alias_field, the original field no longer exist.
Double quote " table alias
FROM (
-- ...
) AS "bigtable"
You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near '"bigtable"' at line N
MySql don't allow the use of " to make table alias (it's technically ok for field alias).
You should use the mysql back quote to escape table alias name, like AS `My Table Alias`
Correct query :
SQL Fiddle
MySQL 5.6 Schema Setup:
CREATE TABLE events
(`eid` int, `start_date` varchar(10), `manager_email` varchar(15))
;
INSERT INTO events
(`eid`, `start_date`, `manager_email`)
VALUES
(1, '2016-11-11', 'mail_1#mail.com'),
(2, '2016-11-12', 'mail_2#mail.com'),
(3, '2016-11-13', 'mail_3#mail.com'),
(4, '2016-11-14', 'mail_4#mail.com'),
(5, '2016-11-15', 'mail_5#mail.com'),
(6, '2016-11-16', 'mail_6#mail.com'),
(7, '2016-11-17', 'mail_7#mail.com')
;
CREATE TABLE attendee_survey_results
(`id` int, `swoop_event_id` int, `startup_weekend_city` varchar(6), `attendee_nps` int)
;
INSERT INTO attendee_survey_results
(`id`, `swoop_event_id`, `startup_weekend_city`, `attendee_nps`)
VALUES
(1, 1, 'city_1', 1),
(2, 2, 'city_2', 22),
(3, 3, 'city_3', 3),
(4, 1, 'city_4', 4),
(5, 2, 'city_5', 5),
(6, 3, 'city_6', 9),
(7, 7, 'city_7', 17)
;
Query 1:
SELECT
AVG(`bigtable`.`attendee_nps`)
FROM
(
SELECT DISTINCT
`asr`.`swoop_event_id` AS `Swoop ID`,
`asr`.`startup_weekend_city` AS `SW City`,
`swooptable`.`start_date` AS `date`,
`gizmo`.`attendee_nps` AS `attendee_nps`
FROM `attendee_survey_results` AS `asr`
JOIN
(
SELECT
`swoop_event_id`,
(
100 * count(CASE WHEN `attendee_nps` >= 9 THEN 1 END)
/ count(`attendee_nps`)
- 100 * count(CASE WHEN `attendee_nps` <= 6 THEN 1 END)
/ count(`attendee_nps`)
) AS `attendee_nps`
FROM `attendee_survey_results`
GROUP BY `swoop_event_id`
) AS `gizmo`
ON `gizmo`.`swoop_event_id` = `asr`.`swoop_event_id`
JOIN
(
SELECT `eid`, `start_date`, `manager_email`
FROM `events`
) AS `swooptable`
ON `gizmo`.`swoop_event_id` = `swooptable`.`eid`
) AS `bigtable`
Results:
| AVG(`bigtable`.`attendee_nps`) |
|--------------------------------|
| -14.28571429 |
I'm trying to make a query but I don't know how to do it. I'm a newbie in SQL querying, so please be patient.
Here is what I have:
select
count(*), sum(time), 'peter'
from
(select *
from ACTUATION
where OPERATOR != 'peter'
and team in (select eq1.name
from TEAMWORKS eq1
where operator1 = 'peter'
or operator2 = 'peter'
or operator3 = 'peter'
or operator4 = 'peter'
or operator5 = 'peter'
or operator6 = 'peter'
or operator7 = 'peter' )) as a
This works OK, but I need to do it for every operator in the table operators and don't know how to do it.
I've tried several subqueries and groups by, but I can not make it work properly.
Kind regards
EDIT:
Let's say I have this DataBase:
create table TEAMWORKS
(
operator1 varchar(50),
operator2 varchar(50),
operator3 varchar(50),
operator4 varchar(50),
operator5 varchar(50),
operator6 varchar(50),
operator7 varchar(50),
name varchar(50)
)
insert into TEAMWORKS (operator1,operator2,name) values ('Peter', 'Paul', 'Pe-Pa')
insert into TEAMWORKS (operator2,operator3,name) values ('Peter', 'John', 'Pe-Jo')
insert into TEAMWORKS (operator1,operator4,name) values ('John', 'Paul', 'Jo-Pa')
insert into TEAMWORKS (operator5,operator6,name) values ('John', 'Peter', 'Jo-Pe')
create table OPERATORS
(
name varchar(50),
surname varchar(50)
)
insert into OPERATORS (name,surname) values ('Peter', 'Font')
insert into OPERATORS (name,surname) values ('Paul', 'Bridges')
insert into OPERATORS (name,surname) values ('John', 'Oldfield')
create table ACTUATION
(
ID int,
time int,
operator varchar(50),
team varchar(50),
description varchar(999)
)
insert into ACTUATION (ID,time,operator,team,description) values (1,30,'Peter', '','Pick flowers')
insert into ACTUATION (ID,time,operator,team,description) values (2,15,'Paul', '','Throw flowers')
insert into ACTUATION (ID,time,operator,team,description) values (3,30,'Peter', 'Jo-Pe','Pick stones')
insert into ACTUATION (ID,time,operator,team,description) values (4,5,'John', 'Jo-Pe','Throw stones')
insert into ACTUATION (ID,time,operator,team,description) values (5,15,'Paul', 'Jo-Pa','Throw tables')
insert into ACTUATION (ID,time,operator,team,description) values (6,30,'Peter', 'Pe-Pa','Pick tables')
And I need to get the time used by Every operator in table OPERATORS where he is not the main operator in Table ACTUATIONS, but part of a TEAM in the ACTUATIONS.
In the given example, I would like to get as result:
Operator #Actuations (count) Time(sum) 'Origin (only for clarifying, list not needed)
Peter 1 5 'from actuation #4
Paul 1 30 'from actuation #6
John 2 45 'form actuations #3 and #5
Hope now it's more clear
EDIT 2:
You have a working example here: http://sqlfiddle.com/#!3/91373/6/0
That's the desired result, but obviously I would not be able to use unions as I would not know the Operators.
Edit:
After OP having altered the question, here is a working query for your example:
select top 1000
o.name
, count(1)
, sum(a.time)
from OPERATORS o
left join (
select
unpvt.Operator
, unpvt.ColName
, unpvt.name
from
(select name, operator1, operator2, operator3, operator4, operator5, operator6, operator7
from TEAMWORKS) ot
UNPIVOT (
Operator FOR ColName IN (operator1, operator2, operator3, operator4, operator5, operator6, operator7)
) as unpvt
) tw on o.name = tw.Operator
left join ACTUATION a on a.team = tw.name
where a.operator != o.name
group by o.name
This returns your decired result :)
Old Post:
I don't know the structure of the database, but I've had a go at guessing.
Firstly, i create to temporary tables in T-SQL - this is done for testing.
I've previously been down-voted for using temporary tables without descriptions in answers on SO for being to complicated, so please ask if you don't understand
declare #ACTUATION table(
time datetime default getdate()
, Operator varchar(8)
, Workteam varchar(8)
)
insert into #ACTUATION (Operator, Workteam)
select 'Gunnar', 'Peter'
union all select 'Peter', 'Gunnar'
We also need the table Teamworks:
declare #TEAMWORKS table(
time datetime default getdate()
, sometext varchar(32)
, operator1 varchar(8)
, operator2 varchar(8)
, operator3 varchar(8)
, operator4 varchar(8)
, operator5 varchar(8)
, operator6 varchar(8)
, operator7 varchar(8)
)
insert into #TEAMWORKS(sometext, operator1, operator2, operator3, operator4, operator5, operator6, operator7)
select 'Blah Blah', 'Gunnar', 'Jack', 'Sam', 'Joe', 'Lee', 'Jane', 'Jim'
union all select 'More Blah', 'Bob', 'Sal', 'Phil', 'Clark', 'Jones', 'Sue', 'Peter'
union all select 'Even more Blaah', 'Im', 'Running', 'out of', 'dummy', 'names', 'Peter', 'Gunnar'
What I've done here is to create two "virtual" tables, which only consist as variables. Eg. they are recreated each time the query is run. However, I think that they are a great tool when trying to visualize.
Next step is to UNPIVOT the operators to a more normalized structure.
This is done using the following query:
select
unpvt.time
, unpvt.sometext
, unpvt.Operator
, unpvt.ColName
from
(select time, sometext, operator1, operator2, operator3, operator4, operator5, operator6, operator7
from #TEAMWORKS) ot
UNPIVOT (
Operator FOR ColName IN (operator1, operator2, operator3, operator4, operator5, operator6, operator7)
) as unpvt;
This query should make a great starting point for further joins.
For instance, I assume that this is (partly) the result you want:
select
a.Operator
, count(1)
, max(t.time)
from #ACTUATION a
left join (
select
unpvt.Operator
, unpvt.time
from (
select time, sometext, operator1, operator2, operator3, operator4, operator5, operator6, operator7
from #TEAMWORKS
) ot
UNPIVOT (
Operator FOR ColName IN (operator1, operator2, operator3, operator4, operator5, operator6, operator7)
) as unpvt
) t on a.Workteam = t.Operator
where a.Operator != t.Operator
group by a.Operator