i have a table
Id
Month
Salary
1
1
20
2
1
20
1
2
30
2
2
30
3
2
40
1
3
40
3
3
60
1
4
60
3
4
70
I was trying to remove some max month in each id . I was trying the following query
select * from Employee
where id , month not in ( select distinct id, max(Month) over(partition by id ) from Employee)
I cant understand what wrong with this query why cant i do this way. Is there any alternative way for this
Your method should work. The syntax is:
select *
from Employee
where (id, month) not in (select distinct id, max(Month) over(partition by id )
from Employee
)
I wouldn't recommend this approach. The window function is superfluous, when you just want aggregation:
select *
from Employee
where (id, month) not in (select id, max(Month)
from Employee
group by id
);
Or a correlated subquery seems more natural to me:
select e.*
from Employee e
where month < (select max(e2.Month)
from Employee e2
where e2.id = e.id
);
This has the advantage that it can use an index on Employee(id, month) and is probably the best performing way to write the query.
Related
I have 2 tables in Mysql. I want to regroup and count the Number of Orderid per month for each customer. If there is no order, I would like to add 0.
Customer Table
CustomerID
1
2
3
Order Table
OrderId CustomerID Date
1 1 2022-01-02
2 1 2022-01-04
3 2 2022-02-03
4 2 2022-03-03
Expect results
CustomerID Date CountOrderID
1 2022-01 2
2 2022-01 1
3 2022-01 0
1 2022-02 0
2 2022-02 1
3 2022-02 0
1 2022-03 0
2 2022-03 1
3 2022-03 0
How I can do this in Mysql?
SELECT customer.CustomerID,
year_month.y_m AS `Date`,
COUNT(order.OrderId) AS CountOrderID
FROM customer
CROSS JOIN (
SELECT DISTINCT DATE_FORMAT(`date`, '%Y-%m') AS y_m
FROM order
) AS year_month
LEFT JOIN order ON order.CustomerID = customer.CustomerID
AND DATE_FORMAT(order.`date`, '%Y-%m') = year_month.y_m
GROUP BY 1, 2;
If order table does not contains for some year and month then according row won't present in the output. If you need in it then you'd generate calendar table instead of year_month subquery.
you can reduce the number of cte's I added more here to explain the steps:
first you need the format year and month, for that I used DATE_FORMAT() function
since you need to have all the combination of dates and the year month you need a cross join. This will produce all the distinct dates with all the distinct customer id's. In other words all the pairs between dates and customer id
once you have a table with all the combinations you need to pass the actual data with the left join this will produce null where you actually don't have rows and hence will produce 0 when the count is performed
the last step is simply count function
with main as (
select distinct DATE_FORMAT(date,'%Y-%m') as year_month from order
),
calendar as (
select * from customer
cross join main
),
joining_all as (
select
calendar.*,
order. OrderId
left join order
on calendar.CustomerID = order.CustomerID
and calendar.year_month = DATE_FORMAT(order.date,'%Y-%m')
)
select
CustomerID,
year_month as Date,
count(OrderId) as CountOrderID
from joining_all
group by 1,2
maybe the shorter version can work with the code below. if runs into syntax you can use the one above
with main as (
select distinct DATE_FORMAT(date,'%Y-%m') as year_month from order
cross join customer
)
select
main.CustomerID,
main.year_month as Date,
count(order.OrderId) as CountOrderID
from main
left join order
on main.CustomerID = order.CustomerID
and main.year_month = DATE_FORMAT(order.date,'%Y-%m')
group by 1,2
Here's my "customers" table:
To get number of enquiries per for a particular month and year, I'm using following query:
SELECT YEAR(customer_date) AS Year, MONTH(customer_date) AS Month, COUNT(customer_id) AS Count FROM customers WHERE customer_product = 6 GROUP BY YEAR(customer_date), MONTH(customer_date)
I get following result:
You can see that as there is no enquery in the April month, so no row fetched for month number 4. But I want 0 value in Count column if there is no record found in that particular month and year.
This is what I want:
One option uses a calendar table to represent all months and years, even those which do not appear in your data set:
SELECT
t1.year,
t2.month,
COUNT(c.customer_id) AS Count
FROM
(
SELECT 2017 AS year UNION ALL
SELECT 2018
) t1
CROSS JOIN
(
SELECT 1 AS month UNION ALL
SELECT 2 UNION ALL
SELECT 3 UNION ALL
SELECT 4 UNION ALL
SELECT 5 UNION ALL
SELECT 6 UNION ALL
SELECT 7 UNION ALL
SELECT 8 UNION ALL
SELECT 9 UNION ALL
SELECT 10 UNION ALL
SELECT 11 UNION ALL
SELECT 12
) t2
LEFT JOIN customers c
ON t1.year = YEAR(c.customer_date) AND
t2.month = MONTH(c.customer_date)
WHERE
c.customer_product = 6
GROUP BY
t1.year,
t2.month
ORDER BY
t1.year,
t2.month;
Note: The above query can probably be made faster by actually creating dedicated calendar tables in your MySQL schema.
The following index on the customers table might help:
CREATE INDEX idx ON customers(customer_product, customer_id);
This might make the join between the calendar tables and customers faster, assuming that the customer_product = 6 condition is restrictive.
I have an SQL table that stores reports. Each row has a customer_id and a building_id and when I have the customer_id, I need to select the latest row (most recent create_date) for each building with that customer_id.
report_id customer_id building_id create_date
1 1 4 1553561789
2 2 5 1553561958
3 1 4 1553561999
4 2 5 1553562108
5 3 7 1553562755
6 3 8 1553570000
I would expect to get report id's 3, 4, 5 and 6 back.
How do I query this? I have tried a few sub-selects and group by and not gotten it to work.
If you are using MySQL 8+, then ROW_NUMBER is a good approach here:
WITH cte AS (
SELECT *, ROW_NUMBER() OVER (PARTITION BY customer_id, building_id
ORDER BY create_date DESC) rn
FROM yourTable
)
SELECT
report_id,
customer_id,
building_id,
create_date
FROM cte
WHERE rn = 1;
If there could be more than one customer/building pair tied for the latest creation date, and you want to capture all ties, then replace ROW_NUMBER with RANK, and use the same query.
Another variation:
SELECT a.*
FROM myTable a
WHERE a.create_date = (SELECT MAX(create_date)
FROM myTable b
WHERE b.customer_id = a.customer_id
AND b.building_id = a.building_id)
Can try doing a search for "effective dated records" to see various approaches.
I have a table with following format -
Customer_id Purchase_date
c1 2015-01-11
c2 2015-02-12
c3 2015-11-12
c1 2016-01-01
c2 2016-12-29
c4 2016-11-28
c4 2015-03-15
... ...
The table essentially contains customer_id with their purchase_date. The customer_id is repetitive based on the purchase made on purchase_date. The above is just a sample data and the table contains about 100,000 records.
Is there a way to partition the customer based on pre-defined category data
Category Partitioning
- Category-1: Customer who has not made purchase in last 10 weeks, but made a purchase before that
- Category-2: Customer who as not made a purchase in last 5 weeks, but made purchase before that
- Category-3: Customer who has made one or more purchase in last 4 weeks or it has been 8 weeks since the first purchase
- Category-4: Customer who has made only one purchase in the last 1 week
- Category-5: Customer who has made only one purchase
What I'm looking for is a query that tells customer and their category -
Customer_id Category
C1 Category-1
... ...
The query can adhere to - oracle, postgres, sqlserver
From your question it seems that a customer can fall in multiple categories. So lets find out the customers in each category and then take UNION of the results.
SELECT DISTINCT Customer_Id, 'CATEGORY-1' AS Category FROM mytable GROUP BY
Customer_Id HAVING DATEDIFF(ww,MAX(Purchase_date),GETDATE()) > 10
UNION
SELECT DISTINCT Customer_Id, 'CATEGORY-2' AS Category FROM mytable GROUP BY
Customer_Id HAVING DATEDIFF(ww,MAX(Purchase_date),GETDATE()) > 5
UNION
SELECT DISTINCT Customer_Id, 'CATEGORY-3' AS Category FROM mytable GROUP BY
Customer_Id HAVING DATEDIFF(ww,MAX(Purchase_date),GETDATE()) < 4 OR
DATEDIFF(ww,MIN(Purchase_date),GETDATE()) =8
UNION
SELECT DISTINCT Customer_Id, 'CATEGORY-4' AS Category FROM mytable WHERE
DATEDIFF(ww,Purchase_date,GETDATE())<=1 GROUP BY Customer_Id having
COUNT(*) =1
UNION
SELECT DISTINCT Customer_Id, 'CATEGORY-5' AS Category FROM mytable GROUP BY
Customer_Id HAVING COUNT(*) =1
ORDER BY Category
Hope this serves your purpose.
Thanks
you can use something like this
with myTab as (
SELECT Customer_id ,MIN(Purchase_date) AS Min_Purchase_date,MAX(Purchase_date) AS Max_Purchase_date
, SUM(CASE WHEN Purchase_date>= DATEADD(WEEk ,-1,GETDATE()) THEN 1 ELSE 0 END ) AS Count_LastWeek
, COUNT(*) AS Count_All
FROM Purchases_Table
GROUP BY Customer_id
)
SELECT Customer_id
, CASE WHEN Max_Purchase_date < DATEADD(WEEK,-10,GETDATE()) THEN 'Category-1'
WHEN Max_Purchase_date < DATEADD(WEEK,-5,GETDATE()) THEN 'Category-2'
WHEN Max_Purchase_date >= DATEADD(WEEK,-4,GETDATE())
OR DATEDIFF(WEEK, Min_Purchase_date,Max_Purchase_date) >= 8 THEN 'Category-3'
WHEN Count_LastWeek = 1 THEN 'Category-4'
WHEN Count_All = 1 THEN 'Category-5'
ELSE 'No Category'
END
FROM myTab
Say i have a table called "users" and it has 40 rows of records, each row has fields:
id, firstname, group_id, login_count, stay_on_page_count
groups has
administrator (1), manager (2), employee (3)
is it possible to create a query that will sort and order the rows this way
group _id stay_on_page_count login_count
========= ================== ===========
1 100mins 100
1 90mins 90
2 100mins 100
3 100mins 100
1 80mins 80
1 70mins 70
2 90mins 90
3 90mins 90
1 60mins 60
1 50mins 50
2 80mins 80
3 80mins 80
1 40mins 40
1 30mins 30
2 70mins 70
3 70mins 70
Basically I would like to create a 4x4 grid view using the query result. the pseudo code is probably
SELECT all FROM user table and to group the result in to cluster of 4,
while each 4 should have ORDER BY group_id ASC as first priority (1,1,2,3)
AND stay_on_page_count ORDER BY DESC as second priority,
AND login_count ORDER BY DESC as last or third priority
i don't know if the pseudo code explains enough, but that's the only thing i can came up with :)
And if its possible, then will it sacrifice performance?
Is there any better approach to accomplish this?
I am using Mysql and PHP (CakePHP 2.x)
Thanks
One approach (using summarised as a table holding the summarised values listed in the question):
select * from
(select s1.*,
#rank1:=#rank1+1 as rankcalc,
floor(#rank1/2) rankgroup,
#rank1%2 rankingroup
from (select #rank1:=1) r1
cross join
(select * from summarised where group_id=1 order by stay_on_page_count desc) s1
union all
select s2.*,
#rank2:=#rank2+1 as rankcalc,
#rank2 rankgroup,
2 rankingroup
from (select #rank2:=0) r2
cross join
(select * from summarised where group_id=2 order by stay_on_page_count desc) s2
union all
select s3.*,
#rank3:=#rank3+1 as rankcalc,
#rank3 rankgroup,
3 rankingroup
from (select #rank3:=0) r3
cross join
(select * from summarised where group_id=3 order by stay_on_page_count desc) s3
) sq order by rankgroup, rankingroup
SQLFiddle here.
Note that this solution is dependent on the ordering being evaluated in the sequence specified in the sub-queries, and not overridden by the optimizer - this should work in current versions of MySQL, but may not work in MariaDB (the open source fork of MySQL) or future versions of MySQL.
While this will generate the order you want it's very much a forced and hard coded effort. It makes several assumptions like login_count is always in the ranges you listed.
SELECT group_ID, Stay_on_Page_count, Login_Count, myset
FROM (
SELECT group_ID, Stay_on_Page_Count, Login_Count, 1 as mySet
FROM USERS
WHERE (GROUP_ID=1 and Login_count >= 90) OR (group_ID in (2,3) and Login_Count=100)
UNION
SELECT group_ID, Stay_on_Page_Count, Login_Count, 2 as mySet
FROM USERS
WHERE (GROUP_ID=1 and Login_count Between 70 and 80)
OR (group_ID in (2,3) and Login_Count=90)
UNION
SELECT group_ID, Stay_on_Page_Count, Login_Count, 3 as mySet
FROM USERS
WHERE (GROUP_ID=1 and Login_count Between 50 and 60)
OR (group_ID in (2,3) and Login_Count=80)
UNION
SELECT group_ID, Stay_on_Page_Count, Login_Count, 4 as mySet
FROM USERS
WHERE (GROUP_ID=1 and Login_count Between 30 and 40)
OR (group_ID in (2,3) and Login_Count=70))
ORDER BY myset, group_ID, Login_count Desc