I thought about it for a while and you either do it like the first post I made for rhetro or you do it with a cursor. The cursor...

OK...just thinking about this a little more. If the above isn't what you had in mind, the next course of action gets trickier if you want to go rhetro-active.

CREATE...

UPDATE myTable

SET myFlag = 1

FROM myTable mt1

WHERE NOT EXISTS

(

SELECT *

FROM myTable mt2

WHERE mt1.marcode

AND DATEPART(yy,mt1.dateTimeStamp) = DATEPART(yy,mt2.dateTimeStamp)

AND...

In a SQL Task..

Select FOO

FROM schema.table2

WHERE...

Click the parameters button...go to the output parameters tab...

pick FOO from the drop-down

Now you have the variable to procede to the datapump. From your...

I was thinking about when I have multiple statements in an SQL task. It would bomb on me. On another hand, large queries are easier to mange in...

create a stored proc on your local server

create procedure sp_GetDataFromOracleTable

@Foo

AS

select *

from schema.table

where id_column > @Foo

When you call it inside the datapump...

EXEC sp_GetDataFromOracleTable ?

click...

great minds think alike

My CIO almost hit the floor when I showed him. That is awesome!

CREATE FUNCTION fn_StripChars(@val varchar(50)) RETURNS int

AS

BEGIN

DECLARE @int varchar(50)

WHILE LEN(@val) > 0

BEGIN

IF ISNUMERIC(LEFT(@val,1)) = 1 SET @int = ISNULL(@int,'') + LEFT(@val,1)

...

Run the first task 5 times on one job (00:00 to 16:01)

Run the last task on 1 job with 2 steps (20:00 once a day)

are you making a Windows or SQL Server Authentication Connection?

are you using named pipes?

are you using a DSN or DSN'less connection?

you may want to Google "SQL Server 80004005"

I just had to do this task a few weeks ago... and a quick little reminder because it's not the first thing that I thought of. You must script...

That's really interesting, but I think recursion is more fun! unless you run out of memory that is (guilty).

I'd keep track of the files you've already uploaded in a table, and move the processed file to a "processed" directory. It's good practice anyhow. Do an inner...

Doesn't task3 wait for task1 and task2 to complete before execution?

you can just connect them sequentially, if you're looking to find an error.