Using sub query result in multiple exist conditions - sql

I'm trying to get all products based on two conditions on another table. See my example db data.
Product
--------------
ProductID ProductName
1 P1
2 P2
3 P3
AuditLog
------------
Event Date ProductId
ApproveProduct 2016-01-27 16:00 1
ViewProduct 2016-01-27 17:00 1
ViewProduct 2016-01-27 15:00 2
ApproveProduct 2016-01-27 17:00 2
Based on the example db data, I would like to get all products that have ViewProduct auditlog entry AFTER ApproveProduct entry based on the Date. I would like to return first product since second product has the AuditLog entries in wrong order:
ProductID ProductName
1 P1
I'm trying to achieve something like below, but obviously that code is not valid:
SELECT p.* FROM Product p
WHERE EXISTS (SELECT TOP 1 Date
FROM AuditLog WHERE Event = 'ApproveProduct' AND ProductId = p.ProductID) ap
AND EXISTS (SELECT 1
FROM AuditLog WHERE Event = 'ViewProduct' AND Date > ap.Date AND ProductId = p.ProductID)

You can use conditional aggregation to get the ProductIDs that you want:
SELECT ProductId
FROM AuditLog
GROUP BY ProductId
HAVING MAX(CASE WHEN Event = 'ViewProduct' THEN Date END) >
MAX(CASE WHEN Event = 'ApproveProduct' THEN Date END)
and also the operator IN to get the product details:
SELECT *
FROM Product
WHERE ProductId IN (
SELECT ProductId
FROM AuditLog
GROUP BY ProductId
HAVING MAX(CASE WHEN Event = 'ViewProduct' THEN Date END) >
MAX(CASE WHEN Event = 'ApproveProduct' THEN Date END)
)
See the demo.
Results:
> ProductID | ProductName
> --------: | :----------
> 1 | P1

The below query should do it
select ProductId
from
(
select
ProductId,
Event,
Lag(Event, 1) over (partition by ProductId order by Date) as PrevEvent
from AuditLog
where Event in ('ApproveProduct', 'ViewProduct')
)
where PrevEvent = 'ApproveProduct' and Event = 'ViewProduct'
I didn't add a join with the Product table to make the SQL clearer but it should be trivial.

Related

How do I get the count for orders with a certain code only when there are multiple codes for the same order possible

Sorry this is my first post - please let me know if something doesn't make sense!
I'm trying to get a count of the number of orders with a specific code XXX ONLY
lets say table A looks something like this
|ORDER ID | ITEM CODE |
123 XXX
123 YYY
123 YYY
456 XXX
456 XXX
456 XXX
789 XXX
000 YYY
what i want in the output is:
order 123 and 000 not to count
and order 456 and 789 to count as 1 each
I only want the count of the unique orders which have item code XXX ONLY
so the count/ output of the final query should be 2
currently what i have is
select order_id, item code, count(order_id) from table a
where item code = 'XXX'
group by order_id, item code
order by count(order_id)
which outputs me the following
ORDER_ID | ITEM CODE | COUNT(ORDER_ID)
123 XXX 1
345 XXX 3
789 XXX 1
This is wrong because I want the output as described above
Thanks in advance!
select order_id
from table_a
group by order_id
having min(item_code) = 'XXX'
and max(item_code) = 'XXX'
Seems like you want this :
select distinct order_id , item_code , 1 as count
from table t1
where not exists (
select 1 from table t2
where t1.order_id = t2.order_id
and t2.item_code <> 'XXX'
)
the count would be always 1 per your question
One option is to use an anti-join. For example:
select distinct t.order_id, 1 as cnt
from table_a t
left join table_a u on u.order_id = t.order_id and u.item_code <> 'XXX'
where t.item_code = 'XXX' and u.order_id is null
Result:
ORDER_ID CNT
--------- ---
789 1
456 1
See running example at db<>fiddle.
EDIT
To get the total count only, tweak the query as shown below:
select count(distinct t.order_id) as cnt
from table_a t
left join table_a u on u.order_id = t.order_id and u.item_code <> 'XXX'
where t.item_code = 'XXX' and u.order_id is null
Result:
CNT
---
2
See running example at db<>fiddle.

how to get value using latest date from one table and joining to another table

i have 1 table inventory_movement here is data in table
product_id | staff_name | status | sum | reference_number
--------------------------------------------------
1 zes cp 1 000122
2 shan cp 4 000133
i have another table inventory_orderproduct where i have cost date
orderdate product_id cost
--------------------------------
01/11/2018 1 3200
01/11/2018 2 100
02/11/2018 1 4000
02/11/2018 1 500
03/11/2018 2 2000
i want this result
product_id| staff_name | status | sum reference_number | cost
--------------------------------------------------------------
1 zes cp 1 000122 4000
2 shan cp 4 000133 2000
here is my query
select ipm.product_id,
case when ipm.order_by_id is not null then
(select au.first_name from users_staffuser us inner join auth_user au on us.user_id= au.id
where us.id = ipm.order_by_id) else '0' end as "Staff_name"
,ipm.status,
Sum(ipm.quantity), ip.reference_number
from inventory_productmovement ipm
inner join inventory_product ip on ipm.product_id = ip.id
inner join users_staffuser us on ip.branch_id = us.branch_id
inner join auth_user au on us.user_id = au.id
AND ipm.status = 'CP'
group by ipm.product_id, au.first_name, ipm.status,
ip.reference_number, ip.product_name
order by 1
Here is the solution of your question.its working fine.if you like the answer please vote!
SELECT i.product_id,i.staff_name,i.status,i.sum reference_number ,s.Cost
FROM (SELECT product_id,MAX(cost) AS Cost
FROM inventory_orderproduct
GROUP BY product_id ) s
JOIN inventory_movement i ON i.product_id =s.product_id
In the given situation, this should work fine:
Select table1.product_id, table2.staff_name, table2.status, table2.reference_number,
MAX(table1.cost)
FROM table2
LEFT JOIN table1 ON table1.product_id = table2.product_id
GROUP BY table2.product_id, table2.staff_name, table2.status, table2.reference_number
You can use the below query to get MAX cost for products
SELECT i.product_id,i.staff_name,i.status,i.sum reference_number ,s.MAXCost
FROM (SELECT product_id,MAX(cost) AS MAXCost
FROM inventory_orderproduct
GROUP BY product_id ) s
JOIN inventory_movement i ON i.product_id =s.product_id
For Retrieving the cost using the latest date use the below query
WITH cte as (
SELECT product_id,cost
,ROW_NUMBER() OVER (PARTITION BY product_id ORDER BY orderdate DESC) AS Rno
FROM inventory_orderproduct )
SELECT i.product_id,i.staff_name,i.status,i.sum reference_number ,s.Cost
FROM cte s
JOIN inventory_movement i ON i.product_id =s.product_id
WHERE s.Rno=1
You can use below query it will pick the data according to the latest date
WITH result as (
SELECT product_id,cost
,ROW_NUMBER() OVER (PARTITION BY product_id ORDER BY date DESC)
FROM inventory_orderproduct )
SELECT i.product_id,i.staff_name,i.status,i.sum reference_number ,s.Cost
FROM result s
JOIN inventory_movement i ON i.product_id =s.product_id

Get results based on conditions in other table

I have a table
Orders
ID User
-----------------
1 Matt
2 Chris
3 John
then I have another table
Order_Contact
ID Order_ID Type Timestamp
-------------------------------------------------------------------
1 1 Request 2018-01-01 10:00:00
2 1 Request 2018-01-01 10:35:00
3 1 Response 2018-01-01 11:00:00
4 1 Request 2018-01-01 12:00:00
5 2 Request 2018-01-01 13:00:00
6 2 Response 2018-01-01 14:00:00
My goal is to create a query that returns all Orders that have "open" contact requests that have no response. One response is enough to satisfy multiple requests. Requests that have no response at later datetime than the request are "open". How could I achieve this?
In the above scenario I would like to return
ID User
----------------
1 Matt
First, get the latest entry per order:
SELECT Type, MAX(ID)
FROM Order_Contact
GROUP BY Order_ID
Then, filter the "open" entries:
SELECT *
FROM (
SELECT Type, MAX(ID)
FROM Order_Contact
GROUP BY Order_ID
) as max_id
WHERE max_id.Type = 'Request'
Edit:
The solution above works for my MySQL-Installation, but the following solution works even using sqlfiddle (http://sqlfiddle.com/#!9/92ecd4/12):
SELECT *
FROM (
SELECT *
FROM Order_Contact
WHERE ID IN (
SELECT MAX(ID)
FROM Order_Contact
GROUP BY Order_ID
)
) as Reduced_Order_Contact
WHERE Reduced_Order_Contact.Type = 'Request'
Edit 2:
Towards the extension of the question, adding columns is a simple join:
SELECT Orders.ID, Orders.User
FROM (
SELECT *
FROM Order_Contact
WHERE ID IN (
SELECT MAX(ID)
FROM Order_Contact
GROUP BY Order_ID
)
) as Reduced_Order_Contact
INNER JOIN Orders
ON Reduced_Order_Contact.Order_ID = Orders.ID
WHERE Reduced_Order_Contact.Type = 'Request'
Check this query
select
Orders.ID, Orders.[User]
from (
select
*, max([Timestamp]) over (partition by Order_ID) maxDate
from
Order_Contact
) t
join Orders on t.Order_ID = Orders.ID
where
t.maxDate = t.[Timestamp]
and t.[Type] = 'Request'
Updated query due to changes in the question

sql 2000 select id where multiple row conditions are met

I am struggling to get my head around this sql.
I have a function that returns a list of items associated with a Bill of Materials BOM.
The result of the sql select
SELECT
BOM,
ITEMID,
QTY
FROM boms
WHERE
bom='A'
is
BOM | ITEMID | QTY
A | ITEMB | 1
A | ITEMC | 2
Now using that result set I am looking to query my salestable to find sales where ITEMB and ITEMC were sold in enough quantity.
The format of the salestable is as follows
SELECT
salesid,
itemid,
sum(qtyordered) 'ordered'
FROM salesline
WHERE
itemid='ITEMB'
or itemid='ITEMC'
GROUP BY salesid, itemid
This would give me something like
salesid | itemid | ordered
SO-10000 | ITEMB | 1
SO-10001 | ITEMB | 1
SO-10001 | ITEMC | 1
SO-10002 | ITEMB | 1
SO-10002 | ITEMC | 2
ideally I would like to return only SO-10002 as this is the only sale where all necessary units were sold.
Any suggestions would be appreciated. Ideally one query would be ideal but I am not sure if that is possible. Performance is not a must as this would be run once a week in the early hours of the morning.
EDIT
with the always excellent help, the code is now complete. I have wrapped it all up into a UDF which simply returns the sales for a specified BOM over a specified period of time.
Function is
CREATE FUNCTION [dbo].[BOMSALES] (#bom varchar(20),#startdate datetime, #enddate datetime)
RETURNS TABLE
AS
RETURN(
select count(q.SALESID) SOLD FROM (SELECT s.SALESID
FROM
(
SELECT s.SALESID, ITEMID, SUM(qtyordered) AS SOLD
FROM salesline s inner join SALESTABLE st on st.salesid=s.SALESID
where st.createddate>=#startdate and st.CREATEDDATE<=#enddate and st.salestype=3
GROUP BY s.SALESID, ITEMID
) AS s
JOIN dbo.BOM1 AS b ON b.ITEMID = s.ITEMID AND b.QTY <= s.SOLD
where b.BOM=#bom
GROUP BY s.SALESID
HAVING COUNT(*) = (SELECT COUNT(*) FROM dbo.BOM1 WHERE BOM = #bom)) q
)
This should return all sales with an exact match, i.e. same itemid and same quantity:
SELECT s.salesid
FROM
(
SELECT salesid, itemid, SUM(qtyordered) AS ordered
FROM salesline AS s
GROUP BY salesid, itemid
) AS s
JOIN
boms AS b
ON b.itemid = s.itemid
AND b.QTY = s.ordered
WHERE b.BOM='A'
GROUP BY s.salesid
HAVING COUNT(*) = (SELECT COUNT(*) FROM boms WHERE BOM='A');
If you want to return a sale where the quantity is greater than boms.qty youhave to change the join accordingly:
JOIN
boms AS b
ON b.itemid = s.itemid
AND b.QTY <= s.ordered
Untested...
You can do this aggregation and a having clause:
select salesid
from salesline sl
group by salesid
having sum(case when itemid = 'ITEMB' then 1 else 0 end) > 0 and
sum(case when itemid = 'ITEMA' then 1 else 0 end) > 0;
Each condition in the having clause is counting the number of rows with each item.
I think this may get you the results you need. You'll have to replace #BOM with your bom:
SELECT
DISTINCT salesid
FROM
salesline sl
INNER JOIN boms b ON
b.bom = #BOM
AND b.itemid = sl.itemid
GROUP BY salesid, itemid
HAVING SUM(qtyordered) >= b.qty
From what I gather, the first query is used to get the thresholds for returning qualifying sales? Based on your example data rows I assumed that there will only be one line per salesid + itemid (basically acting as a dual field primary key) in the salesline table? If that is true I don't think there is a need to do a SUM as you have in your second example query. Let me know if I'm mistaken in any of my assumptions and I'll adjust my answer.

check if the column value exists in subquery

i have 3 tables Product Category and ProductCategory.
Product table:
ProductID ProductName
1 P1
2 P2
3 P3
Category table:
CategoryID CategoryName
1 C1
2 C2
3 C3
ProductCategory:
ProductID CategoryID
1 1
1 2
1 3
2 3
3 1
3 2
I need a query which returns products which fall under more than 1 categories. Based on the table data above the result would be:
ProductID ProductName
1 P1
3 P3
So i wrote a query to fetch all the ProductID's which have more than one CategoryID's as below:
select ProductID,count(CategoryID)
from ProductCategory
group by Productid
having count(CategoryID)>1)
But when i try to display product details using the below query i get an error:
select *
from Product
where ProductID in (
select ProductID,count(CategoryID)
from ProductCategory
group by Productid
having count(CategoryID)>1))
Is my query wrong? How do i get the required product details which fall in more than one categories?
Remove the COUNT() in the subquery. The result of the subquery when used on IN clause must have only one returned column.
SELECT *
FROM Product
WHERE ProductID IN
(
SELECT ProductID
FROM ProductCategory
GROUP BY Productid
HAVING count(CategoryID) > 1
)
SQLFiddle Demo
or by using JOIN
SELECT a.*
FROM Product a
INNER JOIN
(
SELECT ProductID
FROM ProductCategory
GROUP BY Productid
HAVING count(CategoryID) > 1
) b ON a.ProductID = b.ProductID
SQLFiddle Demo
You can try use CROSS APPLY Operator in SQL Server
SELECT DISTINCT C.ProductID,C.ProductName,A.CategoryID,A.Total
FROM Product C
CROSS APPLY (
Select CA.CategoryID,Total=COUNT(*)
From ProductCategory CA
Where C.ProductID=CA.ProductID
Group By CA.CategoryID Having COUNT(*)>1
) AS A
ORDER BY A.Total DESC
Take a look: http://explainextended.com/2009/07/16/inner-join-vs-cross-apply/