select Tf.*
from SalesOrder SO
join TransportOrder Tf on Tf.SalesOrderID = SO.SalesOrderID
join (
select Sz.TradingPartner, Sz.ExternalSalesOrder, Tz.TransportOrderNumber, max(Tz.Revision) Revision
from SalesOrder Sz
join TransportOrder Tz on Sz.SalesOrderID = Tz.SalesOrderID
group by Sz.TradingPartner, Sz.ExternalSalesOrder, Tz.TransportOrderNumber
) TU on TU.TradingPartner = SO.TradingPartner and TU.ExternalSalesOrder = SO.ExternalSalesOrder and TU.TransportOrderNumber = Tf.TransportOrderNumber and Tf.Revision = TU.Revision
I want to know If I can improve it?
What I want to do:
select the TransportOrders that have the maximum revision.
a transport order can be identified with salesOrder.TradingPartner, salesOrder.ExternalSalerOrder, transportOrder.TransportOrderNumer and transportOrder.Revision (used as a version field)
so I want all the transportorder with last version
You can use the ROW_NUMBER analytical function as follows:
SELECT * FROM
(SELECT TF.*,
ROW_NUMBER()
OVER (PARTITION BY SZ.TRADINGPARTNER,
SZ.EXTERNALSALESORDER,
TZ.TRANSPORTORDERNUMBER
ORDER BY TZ.REVISION DESC) AS RN
FROM SALESORDER SO
JOIN TRANSPORTORDER TF
ON TF.SALESORDERID = SO.SALESORDERID
JOIN SALESORDER SZ
ON SZ.TRADINGPARTNER = SO.TRADINGPARTNER
AND SZ.EXTERNALSALESORDER = SO.EXTERNALSALESORDER
JOIN TRANSPORTORDER TZ
ON SZ.SALESORDERID = TZ.SALESORDERID
AND TRANSPORTORDERNUMBER = TF.TRANSPORTORDERNUMBER
)
WHERE TF.REVISION = TU.REVISION
I would suggest window functions
select st.*
from (select t.*, so.TradingPartner, so.ExternalSalerOrder,
max(t.revision) over (partition by so.TradingPartner, so.ExternalSalerOrder, t.TransportOrderNumber) as max_revision
from SalesOrder SO join
TransportOrder t
on t.SalesOrderID = so.SalesOrderID
) st
where revision = max_revision;
No need to reopen the two tables. You can use window functions like this:
select *
from (
select t.*,
rank() over(
partition by s.TradingPartner, s.ExternalSalerOrder, t.TransportOrderNumer
order by t.Revision desc
) rn
from SalesOrder s
join TransportOrder t on t.SalesOrderID = s.SalesOrderID
) t
where rn = 1
Related
I have a view created in SQL Server Management Studio that brings in certain data, I need to only select the rows with the minimum sequence. For example, in the screenshot see the job number "50773-4", I would only need to see the row with SEQ number 2. I've tried to Group by Min, but to no avail. Any help would be appreciated.
SELECT
TOP (100) PERCENT dbo.Job_Operation.Job,
MIN(dbo.Job_Operation.Sequence) AS SEQ,
dbo.Job_Operation.Work_Center,
dbo.Work_Center.Department
FROM
dbo.Job_Operation
INNER JOIN dbo.Job ON dbo.Job_Operation.Job = dbo.Job.Job
INNER JOIN dbo.User_Values ON dbo.Job.User_Values = dbo.User_Values.User_Values
INNER JOIN dbo.Work_Center ON dbo.Job_Operation.Work_Center = dbo.Work_Center.Work_Center
GROUP BY
dbo.Job_Operation.Job,
dbo.User_Values.Numeric2,
dbo.Work_Center.UVText4,
dbo.Job.Status,
dbo.Job_Operation.Status,
dbo.User_Values.Decimal1,
dbo.Job_Operation.Work_Center,
dbo.Work_Center.Department
HAVING
(dbo.Work_Center.UVText4 = 'Machining')
ORDER BY
dbo.User_Values.Decimal1 DESC,
SEQ
[enter image description here]
I would try the RANK() window function. Perhaps:
SELECT column1,
column2,
rank() OVER (PARTITION BY job ORDER BY seq) AS seq_by_job
Then use this as a nested statement, and filter on only the min rank (i.e. WHERE nested_statement.seq_by_job = 1)
here is one way :
SELECT
TOP (100) PERCENT Job,
Sequence AS SEQ,
Work_Center,
Department
FROM
( select dbo.Job_Operation.Job,
dbo.Job_Operation.Sequence,
dbo.Job_Operation.Work_Center,
dbo.Work_Center.Department,
dbo.User_Values.Decimal1 ,
ROW_NUMBER() over (partition by dbo.Job_Operation.Job,
dbo.User_Values.Numeric2,
dbo.Work_Center.UVText4,
dbo.Job.Status,
dbo.Job_Operation.Status,
dbo.User_Values.Decimal1,
dbo.Job_Operation.Work_Center,
dbo.Work_Center.Department
Order by dbo.Job_Operation.Sequence asc) rn
FROM
dbo.Job_Operation
INNER JOIN dbo.Job ON dbo.Job_Operation.Job = dbo.Job.Job
INNER JOIN dbo.User_Values ON dbo.Job.User_Values = dbo.User_Values.User_Values
INNER JOIN dbo.Work_Center ON dbo.Job_Operation.Work_Center = dbo.Work_Center.Work_Center
) tt
WHERE rn = 1
and UVText4 = 'Machining'
You can do:
with
q as (
SELECT
dbo.Job_Operation.Job,
MIN(dbo.Job_Operation.Sequence) AS SEQ,
dbo.Job_Operation.Work_Center,
dbo.Work_Center.Department,
dbo.User_Values.Decimal1
FROM
dbo.Job_Operation
INNER JOIN dbo.Job ON dbo.Job_Operation.Job = dbo.Job.Job
INNER JOIN dbo.User_Values
ON dbo.Job.User_Values = dbo.User_Values.User_Values
INNER JOIN dbo.Work_Center
ON dbo.Job_Operation.Work_Center = dbo.Work_Center.Work_Center
GROUP BY
dbo.Job_Operation.Job,
dbo.User_Values.Numeric2,
dbo.Work_Center.UVText4,
dbo.Job.Status,
dbo.Job_Operation.Status,
dbo.User_Values.Decimal1,
dbo.Job_Operation.Work_Center,
dbo.Work_Center.Department
HAVING
(dbo.Work_Center.UVText4 = 'Machining')
),
r as (
select *,
row_number() over(partition by job order by seq) as rn
from q
)
select job, seq, work_center, department
from r
where rn = 1
order by Decimal1 DESC
I would like to retrieve one large table of products with the latest rows from all the joined tables via with MAX(ID) of each group (productToken) which ich unique name of the product. Joined tables are - products (store), availability (status), description (products), and price of the product. All of them contains the unique productToken and the mentioned tables can be changed over time by adding a new record (independently) so my aim is to compose one big table (with actual info about the products) via retrieving the lastest record from each table. My code is this. First added product worked well, but things got strange after adding new records to any of the tables (query has retrieved no results).
SELECT *
FROM products
JOIN productsStore ON products.productToken = productStore.productToken
JOIN productsStatus ON products.productToken = productsStatus.productToken
JOIN productsPrice ON products.produstToken = productsPrice.productToken
JOIN categories ON products.categoryToken = categories.categoryToken
WHERE products.shopToken = '$shopToken'
AND products.productID IN
(SELECT MAX(productID)
FROM products
GROUP BY productToken)
AND productsPrice.productPriceID IN
(SELECT MAX(productPriceID)
FROM productsPrice
GROUP BY produktToken)
AND productsStatus.productStatusID IN
(SELECT MAX(productStatusID)
FROM productsStatus
GROUP BY productToken)
AND produktyStore.productStoreID IN
(SELECT MAX(productStoreID)
FROM productsStore
GROUP BY productToken)
AND categories.categoryID IN
(SELECT MAX(categoryID)
FROM categories
GROUP BY categoryToken)
ORDER BY categories.categoryID DESC
I would like to retrieve one large table of products with the latest rows from all the joined tables
I think that you want equality conditions with correlated subqueries in the where clause rather than in conditions with aggregate queries. This lets you filter each joined table with the "latest" record for the given productToken.
SELECT *
FROM products p
JOIN productsStore psr ON psr.productToken = p.productToken
JOIN productsStatus psu ON psu.productToken = p.productToken
JOIN productsPrice ppr ON ppr.produstToken = p.productToken
JOIN categories c ON c.categoryToken = p.categoryToken
WHERE
p.shopToken = '$shopToken'
AND p.productID = (SELECT MAX(p1.productID) FROM products p1 WHERE p1.productToken = p.productToken)
AND psr.productStoreID = (SELECT MAX(psr1.productStoreID) FROM productsStore psr1 WHERE psr1.productToken = p.productToken)
AND psu.productStatusID = (SELECT MAX(psu1.productStatusID) FROM productStatus psu1 WHERE psu1.productToken = p.productToken)
AND ppr.productPriceID = (SELECT MAX(ppr1.productPriceID) FROM productsPrice ppr1 WHERE ppr1.productToken = p.productToken)
AND c.categoryID = (SELECT MAX(c1.categoryID) FROM category c1 WHERE c1.productToken = p.productToken)
If you are running MySQL 8.0 (or MariaDB 10.3 or higher), you can use ROW_NUMBER() in subqueries instead:
SELECT *
FROM (
SELECT p.*, ROW_NUMBER() OVER(PARTITION BY productToken ORDER BY productID DESC) rn
FROM products p
) p
INNER JOIN (
SELECT psr.*, ROW_NUMBER() OVER(PARTITION BY productToken ORDER BY productStoreID DESC) rn
FROM productsStore psr
) psr ON psr.productToken = p.productToken AND psr.rn = 1
INNER JOIN (
SELECT psu.*, ROW_NUMBER() OVER(PARTITION BY productToken ORDER BY productStatusID DESC) rn
FROM productsStatus psu
) psu ON psu.productToken = p.productToken AND psu.rn = 1
INNER JOIN (
SELECT ppr.*, ROW_NUMBER() OVER(PARTITION BY productToken ORDER BY productsPriceID DESC) rn
FROM productsPrice ppr
) ppr ON ppr.productToken = p.productToken AND ppr.rn = 1
INNER JOIN (
SELECT c.*, ROW_NUMBER() OVER(PARTITION BY productToken ORDER BY categoryID DESC) rn
FROM categories c
) c ON c.productToken = p.productToken AND c.rn = 1
WHERE p.shopToken = '$shopToken' AND p.rn = 1
JOIN public.match m ON (s.stadium_id = m.stadium_id)
group
AS (
)
SELECT round_number
,stadium_name
,spectators
FROM (
SELECT round_number
,stadium_name
,spectators
,RANK() OVER (
PARTITION BY round_number ORDER BY spectators DESC
) AS rank1
FROM t1
) AS s1
WHERE rank1 = 1
<br>
Any smaller query than this?
I think you can just use window functions:
select ms.*
from (select m.round_number, s.stadium_name, m.no_spectators,
row_number() over (partition by m.round_number order by m.no_spectators desc) as seqnum
from public.stadium s join
public.match m
on s.stadium_id = m.stadium_id
) ms
where seqnum = 1
order by m.round_number;
I don't see why aggregation would be needed for the inner query.
You can use a subquery to get the max first
select m.round_number, s.stadium_name, MaxSpec
from public.stadium s
JOIN public.match m ON (s.stadium_id = m.stadium_id)
JOIN
(
select m.round_number, MAX(m.no_spectators) as MaxSpec
from public.stadium s
JOIN public.match m ON (s.stadium_id = m.stadium_id)
group by m.round_number
)a on m.no_spectators = a.MaxSpec
Just one more way to skin this cat. Throw your MAX(no_spectators) into a WHERE clause.
SELECT
m.round_number,
s.stadium_name,
m.no_spectators
FROM
PUBLIC.stadium s
JOIN
PUBLIC.match m
ON s.stadium_id = m.stadium_id
WHERE
m.no_spectators = (SELECT MAX(no_spectators) FROM PUBLIC.match);
That should do for an intro class.
The following code gives me multiple lines since there can be more than one Cust_Edit_Log.Edit_Timestamp per Alarm Account. There is no other way for a duplicate to occur. How do I only get the result with the earliest Cust_Edit_Log.Edit_Timestamp date? Thank you in advance for any help you can provide.
Select
AR_Customer.Customer_Number As 'Customer_Number',
AR_Customer.Customer_Name As 'Customer_Name',
AR_Customer_System.Alarm_Account As 'Alarm_Account',
AR_Customer_Site.Address_1 As 'Site_Address_1',
Cust_Edit_Log.UserComments As 'Edit_Log_Cust_User_Comments',
Cust_Edit_Log.Edit_Timestamp As 'Edit_Log_Cust_Timestamp',
Cust_Edit_Log.UserCode As 'Edit_Log_Cust_User'
From
AR_Customer
Inner JOIN AR_Customer_Site On AR_Customer.Customer_Id = AR_Customer_Site.Customer_Id
Left Outer JOIN AR_Customer_System On AR_Customer_Site.Customer_Site_Id = AR_Customer_System.Customer_Site_Id
Left Outer Join CQB_Log_Parse Cust_Edit_Log on AR_Customer.Customer_Id = Cust_Edit_Log.Customer_Id
Where
AR_Customer.Customer_Id <> 1 And
(AR_Customer_System.Alarm_Account Like 'IN%' And
Cust_Edit_Log.UserComments Like 'Edited Customer System IN%')
Order By
AR_Customer.Customer_Number ASC
Use Partition BY:
SELECT
X.*
FROM
(
Select
AR_Customer.Customer_Number As 'Customer_Number',
AR_Customer.Customer_Name As 'Customer_Name',
AR_Customer_System.Alarm_Account As 'Alarm_Account',
AR_Customer_Site.Address_1 As 'Site_Address_1',
Cust_Edit_Log.UserComments As 'Edit_Log_Cust_User_Comments',
Cust_Edit_Log.Edit_Timestamp As 'Edit_Log_Cust_Timestamp',
Cust_Edit_Log.UserCode As 'Edit_Log_Cust_User',
ROW_NUMBER() OVER(Partition BY AR_Customer_System.Alarm_Account,Cust_Edit_Log.Edit_Timestamp ORDER BY AR_Customer_System.Alarm_Account) AS PartNO
From
AR_Customer
Inner JOIN AR_Customer_Site On AR_Customer.Customer_Id = AR_Customer_Site.Customer_Id
Left Outer JOIN AR_Customer_System On AR_Customer_Site.Customer_Site_Id = AR_Customer_System.Customer_Site_Id
Left Outer Join CQB_Log_Parse Cust_Edit_Log on AR_Customer.Customer_Id = Cust_Edit_Log.Customer_Id
Where
AR_Customer.Customer_Id <> 1 And
(AR_Customer_System.Alarm_Account Like 'IN%' And
Cust_Edit_Log.UserComments Like 'Edited Customer System IN%')
)X
WHERE X.PartNo=1
Order By X.Customer_Number ASC
One method uses row_number():
Left Outer Join
(select lp.*,
row_number() over (partition by lp.Customer_Id
order by Edit_Timestamp asc
) as seqnum
from CQB_Log_Parse lp
) Cust_Edit_Log
on AR_Customer.Customer_Id = Cust_Edit_Log.Customer_Id and
seqnum = 1
Maybe try with MIN(Cust_Edit_Log.Edit_Timestamp)
You can try using as below:
;with cte as (
Select
AR_Customer.Customer_Number As 'Customer_Number',
AR_Customer.Customer_Name As 'Customer_Name',
AR_Customer_System.Alarm_Account As 'Alarm_Account',
AR_Customer_Site.Address_1 As 'Site_Address_1',
Cust_Edit_Log.UserComments As 'Edit_Log_Cust_User_Comments',
Cust_Edit_Log.Edit_Timestamp As 'Edit_Log_Cust_Timestamp',
Cust_Edit_Log.UserCode As 'Edit_Log_Cust_User'
,row_number() over(partition by AR_Customer.Customer_Number order by Cust_Edit_Log.Edit_Timestamp) as rownum
From
AR_Customer
Inner JOIN AR_Customer_Site On AR_Customer.Customer_Id = AR_Customer_Site.Customer_Id
Left Outer JOIN AR_Customer_System On AR_Customer_Site.Customer_Site_Id = AR_Customer_System.Customer_Site_Id
Left Outer Join CQB_Log_Parse Cust_Edit_Log on AR_Customer.Customer_Id = Cust_Edit_Log.Customer_Id
Where
AR_Customer.Customer_Id <> 1 And
(AR_Customer_System.Alarm_Account Like 'IN%' And
Cust_Edit_Log.UserComments Like 'Edited Customer System IN%')
--Order By
--AR_Customer.Customer_Number ASC
)
select * from cte where rownum = 1
order by AR_Customer.Customer_Number ASC
It should return some fields from the SystemTable and the LoadStatus column of the latest record in the ProcessHistory table. The relationship is 1 to many:
SELECT ST.[SystemDetailID], ST.[SystemName], LH.LatestLoadStatus
FROM [SystemTable] AS ST
LEFT OUTER JOIN
(
SELECT LHInner.LoadStatus AS LatestLoadStatus, LHInner.SystemDetailID FROM [dbo].[LoadHistory] AS LHInner
WHERE LHInner.LoadHistoryID in
(
SELECT LatestLoadHisotoryID FROM
(
SELECT MAX(LoadHistoryID) as LatestLoadHisotoryID, SystemDetailID FROM [dbo].[LoadHistory]
GROUP BY SystemDetailID
) l
)
) AS LH ON ST.SystemDetailID = LH.SystemDetailID
Thanks,
This is a greatest-n-per-group query.
One Approach
SELECT ST.[SystemDetailID],
ST.[SystemName],
LH.LatestLoadStatus
FROM [SystemTable] AS ST
OUTER APPLY (SELECT TOP 1 *
FROM [dbo].[LoadHistory] LH
WHERE ST.SystemDetailID = LH.SystemDetailID
ORDER BY LoadHistoryID DESC) LH
You can also use row_number
WITH LH
AS (SELECT *,
ROW_NUMBER() OVER (PARTITION BY SystemDetailID
ORDER BY LoadHistoryID DESC) RN
FROM [dbo].[LoadHistory])
SELECT ST.[SystemDetailID],
ST.[SystemName],
LH.LatestLoadStatus
FROM [SystemTable] AS ST
LEFT JOIN LH
ON LH.SystemDetailID = ST.SystemDetailID
AND LH.RN = 1
SELECT ST.[SystemDetailID], ST.[SystemName], LH.LatestLoadStatus
FROM [SystemTable] AS ST
INNER JOIN [dbo].[LoadHistory] AS LH
ON ST.SystemDetailID = LH.SystemDetailID
AND LH.LoadHistoryID IN
(SELECT MAX(LoadHistoryID) as LoadHistoryID
FROM [dbo].[LoadHistory]
GROUP BY SystemDetailID )