This is kind of a double question, just thinking of ways to accomplish my problem.
Also, I'm pretty new to DB2 and stored procedures, so bear with me.
I'm creating a stored procedure that gets a value from two tables using a Left Join statement. This will result in some of the values in the second table returning a null value (since they don't exist in tableB).
DECLARE CURSOR C1 WITH RETURN FOR
select a.name, a.title, b.order from tableA a
left outer join tableB b on a.name = b.name;
Now, I need some way to set these null values to a default value of 0.
The program I'm working with can do it ( CAST-IRON ) but if the result set is too large, it slows down the orchestrations and truncates the job log. So I'm trying to figure it out using the stored procedure.
My first thought was to use the FETCH INTO statement and a WHILE loop.
WHILE AT_END = 0 DO
FETCH C1 INTO CHNAME, CHTITLE, CHORDER;
IF CHORDER IS NULL
THEN SET CHORDER = 0;
END IF;
IF SQLCODE = 100
THEN SET AT_END = 1;
END IF;
END WHILE;
But it seems like that would require a temporary table being created, and declaring another cursor with that table, using an insert command after the 'FETCH INTO'. So I was wondering if there were another way to do this, or to automatically set a default in the select statement?
Set a default in the select statement using COALESCE.
DECLARE CURSOR C1 WITH RETURN FOR
select a.name, a.title, COALESCE(b.order,0) as order
from tableA a
left outer join tableB b on a.name = b.name;
Related
Is it possible to define a with clause in a stored procedure and use it in if else statements because I always get an error?
BEGIN
WITH Test (F, A) AS
(
SELECT FM.ID, FM.Name
FROM [Test.Abc] FM
INNER JOIN [Organization] O on O.ABCID = FM.ID
)
IF(#var = 'case1')
BEGIN
SELECT *
FROM Test F
WHERE NOT F.ID = 'someID'
END
I always get an "Incorrect syntax" error before the if statement
If I move the with clause into the if statement it works fine. But I need the with statement outside to reuse it in different if else statements.
Here's another version of the same answers you're getting:
Your with common table expresson has to be in the same statement as the query that calls it, and it has to be referenced by a query (or other cte) or it is a syntax error.
Reference the documentation Guidelines for Creating and Using Common Table Expressions.
BEGIN -- doing stuff
-- .... doing stuff over here
IF(#var = 'case1')
BEGIN
with Test (F, A) as (
select FM.ID, FM.Name from [Test.Abc] FM
inner join [Organization] O on O.ABCID = FM.ID
)
select * from Test F
where not F.ID = 'someID'
END
-- .... and doing some other stuff over here too
END -- done with this stuff
Just use a temporary table or table variable. The scoping rules of SQL Server ensure that such a table is dropped at the end of the procedure:
BEGIN
select FM.ID, FM.Name
into #test
from [Test.Abc] FM inner join
[Organization] O
on O.ABCID = FM.ID;
IF(#var = 'case1')
BEGIN
select *
from #Test F
where not F.ID = 'someID'
END;
This has the advantage that you can add indexes to the table, and these might improve performance.
WITH is not a standalone, it always a part of a whole statement and only one statement.
It is not recognizable outside the scope ofits statement.
BEGIN
with my_cte (n) as (select 1+1)
select * from my_cte
-- The following statement yields the error "Invalid object name 'my_cte'."
-- select * from my_cte
END
I have two databases running on MSSQL 2005, SOURCE and DESTINATION, which have the same structure and tables in them.
I'm trying to update data from s to d.
In this example, I'm trying to copy data from s to d using a join and only bringing across entries which aren't already in d.
I'm then trying to update the same records just inserted with vales thus:
INSERT DESTINATION.ITEM_REPLENISH_VENDOR ([ITEM_CODE],[VEND_CODE],[PRIMARY_VENDOR],[PURCHASE_MEASURE],[STD_COST],[LAST_COST],[EOQ],[VENDOR_PART_NO],[LEAD_TIME],[COST])
SELECT s.[ITEM_CODE],s.[VEND_CODE],s.[PRIMARY_VENDOR],s.[PURCHASE_MEASURE],s.[STD_COST],s.[LAST_COST],s.[EOQ],s.[VENDOR_PART_NO],s.[LEAD_TIME], s.[COST] FROM SOURCE.dbo.ITEM_REPLENISH_VENDOR s
LEFT OUTER JOIN DESTINATION.dbo.ITEM_REPLENISH_VENDOR d ON (d.ITEM_CODE = s.ITEM_CODE)
WHERE d.ITEM_CODE IS NULL
UPDATE DESTINATION.dbo.ITEM_REPLENISH_VENDOR
SET VEND_CODE='100004', PRIMARY_VENDOR='T',STD_COST='0',LAST_COST='0',COST='0'
WHERE
My issue is once I reach the second WHERE I don't know how to refer to the data I've just updated. This script is going to run either every day at a set time and I don't want to overwrite that whole column with those values, just the entries that have been inserted on this execution.
It looks like you want the output clause This will let you stash away the inserted values.
-- item_code needs to have the same type as the source table
declare #inserted table (item_code int not null primary key);
insert destination.item_replenish_vendor (
[item_code], [vend_code], [primary_vendor],
[purchase_measure], [std_cost], [last_cost],
[eoq], [vendor_part_no], [lead_time],[cost]
) -- save inserted values
output
inserted.item_code into #inserted
select
s.[item_code], s.[vend_code], s.[primary_vendor],
s.[purchase_measure], s.[std_cost], s.[last_cost],
s.[eoq], s.[vendor_part_no], s.[lead_time], s.[cost]
from
source.dbo.item_replenish_vendor s
left outer join
destination.dbo.item_replenish_vendor d
on d.item_code = s.item_code
where
d.item_code is null;
update
d
set
vend_code = '100004',
primary_vendor = 'T',
std_cost = '0',
last_cost = '0,
cost = '0'
from
destination.dbo.item_replenish_vendor d
inner join
#inserted i
on d.item_code = i.item_code;
In this case, you could just put constant values in the insert statement, instead of doing things in two steps...
In the example you have:
UPDATE DESTINATION.dbo.ITEM_REPLENISH_VENDOR
SET VEND_CODE='100004', PRIMARY_VENDOR='T',STD_COST='0',LAST_COST='0',COST='0'
If your VEND_CODE, PRIMARY_VENDOR, STD_COST, LAST_COST, COST are always going to be a static value, you could just put them into the first query.
INSERT DESTINATION.ITEM_REPLENISH_VENDOR ([ITEM_CODE],[VEND_CODE],[PRIMARY_VENDOR],[PURCHASE_MEASURE],[STD_COST],[LAST_COST],[EOQ],[VENDOR_PART_NO],[LEAD_TIME],[COST])
SELECT s.[ITEM_CODE],'100004','T',s.[PURCHASE_MEASURE],'0','0',s.[EOQ],s.[VENDOR_PART_NO],s.[LEAD_TIME], '0'
FROM SOURCE.dbo.ITEM_REPLENISH_VENDOR s
LEFT OUTER JOIN DESTINATION.dbo.ITEM_REPLENISH_VENDOR d ON (d.ITEM_CODE = s.ITEM_CODE)
WHERE d.ITEM_CODE IS NULL
but if they do need to be calculated after insert, then I agree with Laurence's approach.
I'm running the below statement and can't understand why the update affects 0 rows while a select with the same join and where return 1 row. They are both of type Varchar and crmnumberAC is bigger so should be able to take the string from AccountNumber. Thoughts?
begin transaction
update c
set c.crmnumberAC = a.AccountNumber
--select a.name, a.AccountNumber, c.fullname, c.crmnumberAC
from Contact as c
right join Account as a
on c.PFH_Mapping_Ac_ContacId = a.AccountId
WHERE (a.AccountNumber IS NOT NULL AND c.crmnumberAC IS NULL)
OR a.AccountNumber != c.crmnumberAC
rollback transaction
This is the result set when I uncomment the select and just run that as far as the where;
As you can see, your SELECT doesn't return anything from the table Contact, the table you are trying to UPDATE.
The SELECT returns a row just because you have a right join, but in table Contact you have no row that fulfill the where condition.
i am trying to select the last record value from my database by using stored procedure, to do this i set my #UPID parameter as SCORE_IDENTITY(), but there are no output result as all after i execute my stored procedure
CREATE PROCEDURE [dbo].[spAuditLogSelect_NewUser]
#UPID int
AS
BEGIN
SET #UPID = SCOPE_IDENTITY()
SELECT
siUserProfile.UPID
,siUserProfile.ProfileType, siProfileType.RGDName AS ProfileTypeName
,siUserProfile.CBID, siCompany.ComName + ' - ' + siComBranch.ComBranchName AS CBName
,siUserProfile.FullName
,siUserProfile.ShortName
,siUserProfile.SerialCode
,siUserProfile.Serial
,siUserProfile.Gender
from siUserProfile WITH (NOLOCK)
inner join siUserProfileDetail WITH (NOLOCK) on siUserProfile.upid = siUserProfileDetail.UPID
left outer join siReferenceGroupDetail siProfileType WITH (NOLOCK) ON siUserProfile.ProfileType = siProfileType.RGDID
left outer join siComBranch WITH (NOLOCK) on siComBranch.CBID = siUserProfile.CBID
left outer join siCompany WITH (NOLOCK) ON siComBranch.CompanyID = siCompany.CompanyID
where siUserProfile.UPID = #UPID
SCOPE_IDENTITY() is meant to be used right after insert. It won't work in a different session.
To retrieve the latest entry, try top 1:
select top 1 *
...
where siUserProfile.UPID = #UPID
order by
siUserProfile.ID desc
You require to use IDENT_CURRENT(‘tablename’).
Please refer below link which illustrate difference between ##IDENTITY,SCOPE_IDENTITY() and IDENT_CURRENT(‘tablename’).
http://blog.sqlauthority.com/2007/03/25/sql-server-identity-vs-scope_identity-vs-ident_current-retrieve-last-inserted-identity-of-record/
Hope this will help you.
I am assuming that you have some primary key(PK) in your table. Write the procedure, in that procedure fire the query
select * from Your_Table where PK_Column in(select max(PK_Column) from Your_Table)
This way you will be able to fetch the latest record from DB. By opening a cursor, you can play with record in your procedure.
My MS SQL Server stored procedure is:
set ANSI_NULLS ON
set QUOTED_IDENTIFIER ON
GO
ALTER PROCEDURE [dbo].[sp_close]
#DATE NVARCHAR(8)
AS
BEGIN
SELECT appointment_datas.appointment_date
,appointment_datas.appointment_no
,costumer_datas.costumer_name
,appointment_datas.appointment_type
,personel_datas.personel_ADI
FROM [LONI].[dbo].appointment_datas
INNER JOIN [LONI].[dbo].costumer_datas ON appointment_datas.appointment_costumer = costumer_datas.costumer_id
INNER JOIN [LONI].[dbo].personel_datas ON appointment_datas.appointment_personel = personel_datas.personel_id
INNER JOIN [GUONU].[dbo].dayend ON appointment_datas.appointment_no <> dayend.appointment_no COLLATE Turkish_CI_AS
WHERE CONVERT(nvarchar(8),appointment_datas.appointment_date,112) = #DATE
END
With this code, I select duplicate rows from the same records. Actually I want to select fields from [LONI].[dbo].appointment_datas but if appointment_no
is not in [GUONU].[dbo].dayend
SELECT DISTINCT removes duplicates in your output.
But your SQL looks wrong. Are you sure you mean to write:
TABLE1.FIELD1 <> TABLE1.FIELD1
This always evaulates to false. I think you may have an error in your SQL, and that might be why you are getting duplicate values. You should rarely use <> in a join clause, and you shouldn't have the same field on both sides.
Maybe you intended:
ON [DB1].[dbo].TABLE1.FIELD1 <> [DB2].[dbo].TABLE1.FIELD1
but this will generate a Cartesian product of all the rows that don't match. I doubt this is what you really mean. Perhaps you want this:
ON [DB1].[dbo].TABLE1.ID = [DB2].[dbo].TABLE1.ID
WHERE[DB1].[dbo].TABLE1.FIELD1 <> [DB2].[dbo].TABLE1.FIELD1
This matches the rows from each database that have the same ID, but differ in a certain column. Notice that the <> is not in the JOIN clause.
--- UPDATE ---
Perhaps you mean to select the results from the two different databases and then union them?
SELECT appointment_datas.appointment_date
,appointment_datas.appointment_no
,costumer_datas.costumer_name
,appointment_datas.appointment_type
,personel_datas.personel_ADI
FROM [LONI].[dbo].appointment_datas
INNER JOIN [LONI].[dbo].costumer_datas ON appointment_datas.appointment_costumer = costumer_datas.costumer_id
INNER JOIN [LONI].[dbo].personel_datas ON appointment_datas.appointment_personel = personel_datas.personel_id
WHERE CONVERT(nvarchar(8),appointment_datas.appointment_date,112)
UNION
SELECT appointment_datas.appointment_date
,appointment_datas.appointment_no
,costumer_datas.costumer_name
,appointment_datas.appointment_type
,personel_datas.personel_ADI
FROM [GUONU].[dbo].appointment_datas
INNER JOIN [GUONU].[dbo].costumer_datas ON appointment_datas.appointment_costumer = costumer_datas.costumer_id
INNER JOIN [GUONU].[dbo].personel_datas ON appointment_datas.appointment_personel = personel_datas.personel_id
WHERE CONVERT(nvarchar(8),appointment_datas.appointment_date,112)
--- SOLUTION ---
Use NOT EXISTS in WHERE clause. Read comments to see why.
The line
INNER JOIN [DB2].[dbo].TABLE1 ON TABLE1.FIELD1 <> TABLE1.FIELD1
makes no sense, you want to rephrase that...
If I understand your question correctly (after your edit)
but if appointment_no is not in
[GUONU].[dbo].dayend
, you actually want a NOT EXISTS subquery:
set ANSI_NULLS ON
set QUOTED_IDENTIFIER ON
GO
ALTER PROCEDURE [dbo].[sp_close]
#DATE NVARCHAR(8)
AS
BEGIN
SELECT appointment_datas.appointment_date
,appointment_datas.appointment_no
,costumer_datas.costumer_name
,appointment_datas.appointment_type
,personel_datas.personel_ADI
FROM [LONI].[dbo].appointment_datas
INNER JOIN [LONI].[dbo].costumer_datas ON appointment_datas.appointment_costumer = costumer_datas.costumer_id
INNER JOIN [LONI].[dbo].personel_datas ON appointment_datas.appointment_personel = personel_datas.personel_id
WHERE CONVERT(nvarchar(8),appointment_datas.appointment_date,112) = #DATE
AND NOT EXISTS (SELECT 'X' FROM [GUONU].[dbo].dayend WHERE dayend.appointment_no = appointment_datas.appointment_no)
END
SELECT DISTINCT TABLE1.FIELD1
,TABLE2.FIELD1
,TABLE1.FIELD3
,TABLE3.FIELD1
FROM ...
NB in some variants you will have to bracket the field list ie
SELECT DISTINCT (TABLE1.FIELD1
,TABLE2.FIELD1
,TABLE1.FIELD3
,TABLE3.FIELD1 ) FROM ...