Service Desk
asked on
I want to update a column VIA insert trigger
I have a SQL query that creates a table, then creates triggers for the table. The triggers need to update a column in the new record from sums and counts in other tables. I'm pretty new to writing queries, so Im a bit confused. I just cannot seem to get the inserted record values to come up. I've tried using variabbles and using the inserted.values from the inserted table.
I need this working tomorrow so any help is appreciated.....
thanks
I need this working tomorrow so any help is appreciated.....
thanks
/* ---------- CREATE THE TABLE --------------------*/
CREATE TABLE [dwANALYZER].[CalData] (
[F1] [smalldatetime] NULL ,
[MasterAgencyID] [varchar] (20) COLLATE Latin1_General_BIN NULL ,
[ProducerID] [varchar] (5) COLLATE Latin1_General_BIN NULL ,
[Issue_State] [nvarchar] (50) COLLATE Latin1_General_BIN NULL ,
[Category] [nvarchar] (50) COLLATE Latin1_General_BIN NULL ,
[Line] [nvarchar] (50) COLLATE Latin1_General_BIN NULL ,
[Form] [varchar] (4) COLLATE Latin1_General_BIN NULL ,
[Origin] [nvarchar] (10) COLLATE Latin1_General_BIN NULL ,
[Originator] [nvarchar] (80) COLLATE Latin1_General_BIN NULL ,
[Policy_Type] [nvarchar] (50) COLLATE Latin1_General_BIN NULL ,
[Quotes] [int] NULL ,
[QuotePremium] [decimal](38, 2) NULL,
[Issues] [int] NULL ,
[IssuePremium] [decimal](38, 2) NULL,
[Renewals] [int] NULL ,
[RenewalPremium] [decimal](38, 2) NULL,
[Cancels] [int] NULL ,
[CancelPremium] [decimal](38, 2) NULL
) ON [PRIMARY]
GO
SET QUOTED_IDENTIFIER ON
GO
SET ANSI_NULLS ON
GO
/*----------------- create triggers here ----------------------*/
CREATE TRIGGER Q_INSERT ON CalData AFTER INSERT
AS
BEGIN
DECLARE @F1 smalldatetime
DECLARE @MasterAgencyID varchar(20)
DECLARE @ProducerID varchar(5)
DECLARE @Issue_State nvarchar(50)
DECLARE @Category nvarchar (50)
DECLARE @Line nvarchar (50)
DECLARE @Form varchar (4)
DECLARE @Origin nvarchar (10)
DECLARE @Originator nvarchar (80)
DECLARE @Policy_Type nvarchar (50)
SELECT
@F1 = [F1],
@MasterAgencyID = [MasterAgencyID],
@ProducerID = [ProducerID],
@Issue_State = [Issue_State],
@Category = [Category],
@Line = [Line],
@Form = [Form],
@Origin = [Origin],
@Originator = [Originator],
@Policy_Type = [Policy_Type] FROM inserted
/* -------ive tried it two ways ------------------------*/
UPDATE CalData
Set [Quotes] = (SELECT COUNT(pqi.POLICY_NUMBER) AS Policies
FROM dbo.tblCalendarDate cal LEFT OUTER JOIN
dbo.A_PolicyQuoteAndIssueWithInspectionsPlusW1 pqi ON CONVERT(char(10), cal.F1, 102) = CONVERT(CHAR(10), pqi.QuoteDate, 102)
GROUP BY cal.F1, pqi.MasterAgencyID, pqi.ProducerID, pqi.Issue_State, pqi.Category, pqi.Line, pqi.Origin, pqi.Originator, pqi.Policy_Type, pqi.Form
HAVING (cal.F1 = inserted.F1) AND (pqi.MasterAgencyID = inserted.MasterAgencyID) AND (pqi.ProducerID = inserted.ProducerID) AND (pqi.Issue_State = inserted.Issue_State)
AND (pqi.Category = inserted.Category) AND (pqi.Line = inserted.Line) AND (pqi.Form = inserted.Form)
AND (pqi.Origin = inserted.Origin) AND (pqi.Originator = inserted.Originator) AND (pqi.Policy_Type = inserted.Policy_Type))
/* -------------------------and this way as well ---------------*/
UPDATE CalData
Set [QuotePremium] = (SELECT SUM(pqi.IssuedPremium) AS Premium
FROM dbo.tblCalendarDate cal LEFT OUTER JOIN
dbo.A_PolicyQuoteAndIssueWithInspectionsPlusW1 pqi ON CONVERT(char(10), cal.F1, 102) = CONVERT(CHAR(10), pqi.QuoteDate, 102)
GROUP BY cal.F1, pqi.MasterAgencyID, pqi.ProducerID, pqi.Issue_State, pqi.Category, pqi.Line, pqi.Origin, pqi.Originator, pqi.Policy_Type, pqi.Form
HAVING (cal.F1 = @F1) AND (pqi.MasterAgencyID = @MasterAgencyID) AND (pqi.ProducerID = @ProducerID) AND (pqi.Issue_State = @Issue_State)
AND (pqi.Category = @Category) AND (pqi.Line = @Line) AND (pqi.Form = @Form)
AND (pqi.Origin = @Origin) AND (pqi.Originator = @Originator) AND (pqi.Policy_Type = @Policy_Type))
end
ASKER
Dont I need the GROUP BY statement in order for the SUM and COUNT to work? (original code snippet line 63 & 73)
The variables come from the newly inserted row. Therefore the AFTER INSERT should supply them, No?
(original snippet line 44-54)
The variables come from the newly inserted row. Therefore the AFTER INSERT should supply them, No?
(original snippet line 44-54)
Replace <PrimaryKey> with the Primary Key or unique column in CallData:
CREATE TRIGGER Q_INSERT ON CalData
AFTER INSERT
AS
BEGIN
UPDATE c
Set Quotes = d.Policies
FROM CalData c
Inner Join (
Select i.<PrimaryKey>,
COUNT(*) Policies
From dbo.tblCalendarDate cal
Inner Join Inserted i On cal.F1 = i.F1
Inner Join dbo.A_PolicyQuoteAndIssueWithInspectionsPlusW1 pqi ON cal.F1 = pqi.QuoteDate
And i.MasterAgencyID = pqi.MasterAgencyID
And i.ProducerID = pqi.ProducerID
And i.Issue_State = pqi.Issue_State
And i.Category = pqi.Category
And i.Line = pqi.Line
And i.Form = pqi.Form
And i.Origin = pqi.Origin
And i.Originator = pqi.Originator
AND i.Policy_Type = pqi.Policy_Type
Group By
i.<PrimaryKey>
) d On c.PrimaryKey = d.<PrimaryKey>
END
This might be easier to read:
CREATE TRIGGER Q_INSERT ON CalData
AFTER INSERT
AS
BEGIN
UPDATE c
Set Quotes = d.Policies
FROM CalData c
Inner Join (
Select i.<PrimaryKey>,
COUNT(*) Policies
From dbo.tblCalendarDate cal
Inner Join Inserted i On cal.F1 = i.F1
Inner Join dbo.A_PolicyQuoteAndIssueW ithInspect ionsPlusW1 pqi ON cal.F1 = pqi.QuoteDate
And i.MasterAgencyID = pqi.MasterAgencyID
And i.ProducerID = pqi.ProducerID
And i.Issue_State = pqi.Issue_State
And i.Category = pqi.Category
And i.Line = pqi.Line
And i.Form = pqi.Form
And i.Origin = pqi.Origin
And i.Originator = pqi.Originator
AND i.Policy_Type = pqi.Policy_Type
Group By
i.<PrimaryKey>
) d On c.PrimaryKey = d.<PrimaryKey>
END
CREATE TRIGGER Q_INSERT ON CalData
AFTER INSERT
AS
BEGIN
UPDATE c
Set Quotes = d.Policies
FROM CalData c
Inner Join (
Select i.<PrimaryKey>,
COUNT(*) Policies
From dbo.tblCalendarDate cal
Inner Join Inserted i On cal.F1 = i.F1
Inner Join dbo.A_PolicyQuoteAndIssueW
And i.MasterAgencyID = pqi.MasterAgencyID
And i.ProducerID = pqi.ProducerID
And i.Issue_State = pqi.Issue_State
And i.Category = pqi.Category
And i.Line = pqi.Line
And i.Form = pqi.Form
And i.Origin = pqi.Origin
And i.Originator = pqi.Originator
AND i.Policy_Type = pqi.Policy_Type
Group By
i.<PrimaryKey>
) d On c.PrimaryKey = d.<PrimaryKey>
END
ASKER
I see that you have used <PrimaryKey> .
Is this a variable or are you telling me to enter some data here?
There is no primary key for this table as many of the matching fields will contain nulls.
Is this a variable or are you telling me to enter some data here?
There is no primary key for this table as many of the matching fields will contain nulls.
You don't need a group by because you're only returing the one value SUM(pqi.IssuedPremium). The where clause will define which row make up the sum.
Yes, since the trigger is Insert After you will get the values when you do the select.
Yes, since the trigger is Insert After you will get the values when you do the select.
>>Is this a variable or are you telling me to enter some data here? <<
Did you not see my comment:
"Replace <PrimaryKey> with the Primary Key or unique column in CallData:"
>>There is no primary key for this table as many of the matching fields will contain nulls.<<
That is too bad. Can you find some combination that is unique or failiing that add a unique column (an IDENTITY column for example)
Here is the problem with your approach:
Let's suppose that your INSERT statement adds 100 rows, then the following local variables will contain just one value (the last one in the Inserted logical table):
SELECT
@F1 = [F1],
@MasterAgencyID = [MasterAgencyID],
@ProducerID = [ProducerID],
@Issue_State = [Issue_State],
@Category = [Category],
@Line = [Line],
@Form = [Form],
@Origin = [Origin],
@Originator = [Originator],
@Policy_Type = [Policy_Type] FROM inserted
The Trigger fires once per statement and NOT once per row inserted.
Did you not see my comment:
"Replace <PrimaryKey> with the Primary Key or unique column in CallData:"
>>There is no primary key for this table as many of the matching fields will contain nulls.<<
That is too bad. Can you find some combination that is unique or failiing that add a unique column (an IDENTITY column for example)
Here is the problem with your approach:
Let's suppose that your INSERT statement adds 100 rows, then the following local variables will contain just one value (the last one in the Inserted logical table):
SELECT
@F1 = [F1],
@MasterAgencyID = [MasterAgencyID],
@ProducerID = [ProducerID],
@Issue_State = [Issue_State],
@Category = [Category],
@Line = [Line],
@Form = [Form],
@Origin = [Origin],
@Originator = [Originator],
@Policy_Type = [Policy_Type] FROM inserted
The Trigger fires once per statement and NOT once per row inserted.
HBHoffman,
>>You don't need a group by because you're only returing the one value SUM(pqi.IssuedPremium).<<
How do you know? See my previous comments
>>Yes, since the trigger is Insert After you will get the values when you do the select.<<
That is only true if a single row is Inserted. Again, read up on how Triggers really work.
>>You don't need a group by because you're only returing the one value SUM(pqi.IssuedPremium).<<
How do you know? See my previous comments
>>Yes, since the trigger is Insert After you will get the values when you do the select.<<
That is only true if a single row is Inserted. Again, read up on how Triggers really work.
ASKER
Ok, so maybe I'm going about this wrong.
I have the table created above that have fields from [F1] to [PolicyType] making the unique key. However they can contain Nulls. Each of the following fields are a COUNT or a SUM based on date JOINS to the tblCalendarDate table. I DO however have views for each of these calculations (Each view contains a COUNT and SUM, 4 views: Quotes, Issues, Cancels and Renewals) .
So maybe instead of a trigger, could I create the table and use a select statement for each of the fields that need calculations?
Here is the view SQL for quotes
I have the table created above that have fields from [F1] to [PolicyType] making the unique key. However they can contain Nulls. Each of the following fields are a COUNT or a SUM based on date JOINS to the tblCalendarDate table. I DO however have views for each of these calculations (Each view contains a COUNT and SUM, 4 views: Quotes, Issues, Cancels and Renewals) .
So maybe instead of a trigger, could I create the table and use a select statement for each of the fields that need calculations?
Here is the view SQL for quotes
SELECT TOP 100 PERCENT cal.F1, pqi.MasterAgencyID, pqi.ProducerID, pqi.Issue_State, pqi.Category, pqi.Line, pqi.Form, pqi.Origin, pqi.Originator,
pqi.Policy_Type, COUNT(pqi.POLICY_NUMBER) AS Policies, SUM(pqi.IssuedPremium) AS Premium
FROM dbo.tblCalendarDate cal LEFT OUTER JOIN
dbo.A_PolicyQuoteAndIssueWithInspectionsPlusW1 pqi ON CONVERT(char(10), cal.F1, 102) = CONVERT(CHAR(10), pqi.QuoteDate, 102)
GROUP BY cal.F1, pqi.MasterAgencyID, pqi.ProducerID, pqi.Issue_State, pqi.Category, pqi.Line, pqi.Origin, pqi.Originator, pqi.Policy_Type, pqi.Form
HAVING (cal.F1 < GETDATE() - 1)
ORDER BY cal.F1 DESC, pqi.MasterAgencyID, pqi.ProducerID
ASKER
BTW
Here is the rest of the query. This is where it inserts the records
Here is the rest of the query. This is where it inserts the records
/*-----------------------------------------------------------------------------*/
INSERT INTO CalData
(F1,
MasterAgencyID ,
ProducerID,
Issue_State,
Category,
Line,
Form,
Origin,
Originator ,
Policy_Type
)
SELECT
cal.F1,
pqi.MasterAgencyID,
pqi.ProducerID,
pqi.Issue_State,
pqi.Category,
pqi.Line,
pqi.Form,
pqi.Origin,
pqi.Originator,
pqi.Policy_Type
FROM
dbo.tblCalendarDate cal LEFT OUTER JOIN
dbo.A_PolicyQuoteAndIssueWithInspectionsPlusW1 pqi
ON CONVERT(char(10), cal.F1, 102) = CONVERT(CHAR(10), pqi.QuoteDate, 102)
GROUP BY cal.F1, pqi.MasterAgencyID, pqi.ProducerID, pqi.Issue_State, pqi.Category, pqi.Line, pqi.Origin, pqi.Originator, pqi.Policy_Type, pqi.Form
HAVING (cal.F1 > GETDATE() - 90)
>>Ok, so maybe I'm going about this wrong.<<
No necessarily. Did you try my code?
>>I have the table created above that have fields from [F1] to [PolicyType] making the unique key. However they can contain Nulls. <<
It would certainly help if you had a primary key. But the code I originally posted should also work.
No necessarily. Did you try my code?
>>I have the table created above that have fields from [F1] to [PolicyType] making the unique key. However they can contain Nulls. <<
It would certainly help if you had a primary key. But the code I originally posted should also work.
ASKER
Is there a way to create the table and insert/replace Spaces instead of NULL. Wont that allow me to create a PRIMARY key?
And I did try your solution but without a primary key, It wont work.
And I did try your solution but without a primary key, It wont work.
>>And I did try your solution but without a primary key, It wont work.<<
How do you know? I did not say you had to ahve one, I said it would help.
How do you know? I did not say you had to ahve one, I said it would help.
ASKER
Onlines 13, 28 and 29 you reference <PrimaryKey> should I just reference the Columns that are unique?
Since there was no Primary key I didnt replace them with anything.
Can you post how you would write it without the primary key reference?
Since there was no Primary key I didnt replace them with anything.
Can you post how you would write it without the primary key reference?
ASKER
One more thing, Is this going to work on all the records?
You said a trigger only works when a single row is inserted. Does a "Select into" qualify as multiple inserts?
I'm back at the office this morning (and overdue obviously). I'll try all of your suggestions and update you.
Thanks for all the time with this.
You said a trigger only works when a single row is inserted. Does a "Select into" qualify as multiple inserts?
I'm back at the office this morning (and overdue obviously). I'll try all of your suggestions and update you.
Thanks for all the time with this.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
OK So I modified the Code to use the unique columns instead of a primary key. it did not work. Attached is zip file of SQL. (2a)
A Second option is that I have views that have already summed the counts and sums that only require a single row be selected to update the inserted record. I tried to replicate the structure you've used but it hangs for a very long time. See Zip (2b)
The view is created by the code I posted 6-7 messages back.
2a-create-caldata-table.zip
2b-create-caldata-table.zip
A Second option is that I have views that have already summed the counts and sums that only require a single row be selected to update the inserted record. I tried to replicate the structure you've used but it hangs for a very long time. See Zip (2b)
The view is created by the code I posted 6-7 messages back.
2a-create-caldata-table.zip
2b-create-caldata-table.zip
ASKER
OK here is the objective:
I have two tables
1) A table of calendar dates (cal), containing all the dates from 1/1/2006 to 12/31/2012
2) a table (PQI) that contains the following:
MasterAgencyID, ProducerID, Issue_State, Category, Line, Form, Origin, Originator, Policy_Type, QuoteDate, IssueDate, EffDate, CancelDate, Premium
I need to create a single table that groups the first 9 fields uniquelly, Counts the records and Sums the premium, for each of the Dates by way of the JOIN to the Calendar table. (therfore 4 updates) Naturally the dates will all be different therefore the Counts and Sums will be different.
Thast why I'm only trying to get the first one to work, figureing I can replicate the code to use the other dates once the first works.
I hope this helps
Thanks again.
I've tried the last code you sent and it cannot find the reference "i"
I have two tables
1) A table of calendar dates (cal), containing all the dates from 1/1/2006 to 12/31/2012
2) a table (PQI) that contains the following:
MasterAgencyID, ProducerID, Issue_State, Category, Line, Form, Origin, Originator, Policy_Type, QuoteDate, IssueDate, EffDate, CancelDate, Premium
I need to create a single table that groups the first 9 fields uniquelly, Counts the records and Sums the premium, for each of the Dates by way of the JOIN to the Calendar table. (therfore 4 updates) Naturally the dates will all be different therefore the Counts and Sums will be different.
Thast why I'm only trying to get the first one to work, figureing I can replicate the code to use the other dates once the first works.
I hope this helps
Thanks again.
I've tried the last code you sent and it cannot find the reference "i"
>>it did not work.<<
Unfortunately that tell me nothing.
I suggest you ask for a refund and re-post. Hopefully someone will step up to the plate.
Good luck.
Unfortunately that tell me nothing.
I suggest you ask for a refund and re-post. Hopefully someone will step up to the plate.
Good luck.
ASKER
It's hard to explain a problem sometimes, so I can see how it's even harder to explain a solution to a problem that hasnt been communicated completely. Even though it wasn't the answer I was looking for, you were thorough and helpful. It helped me understand how triggers work a bit better and, most of all, it helped me realize I was going about it wrong and made me look into other options. I actually got it done with cursors and updates/inserts. So you deserve the points for all your help. Thank you very much.
Also you can test in in Mgmt Studio by executing the select and supplying values for your variables. see further below
Hope this helps
Open in new window