July 14, 2010 at 9:14 pm
Hi all,
I there any way to join two sources with no relation...
I have a Stored procedure to generate a unique ID(custom ids) based on number of rows I fetch from the source, but the problem is that the ID's generated are inserted into a separate table by the stored procedure itself, now I have to find a method to use those ID's in my data flow and join those with my source returned rows, and insert both IDs and other information to my destination....
but as I don't have anything to lookup on I cant join it with data flow rows, also I can join any ID with any row...
This is not happening
Any help on this...
,
------------------
Why not ?
July 15, 2010 at 12:47 am
Please provide sample input data, along with desired output data.
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
July 15, 2010 at 6:19 am
Actually, I dont have any input in a way to be presented, Its just I want an approach to do this problem.
I can put it in steps like this:
> OLE DB source will return 10 rows
> Existing Stored procedure will generate 10 ID's for those 10 rows from the source and put them in a separate table (say tableB) to be picked up from.
> Now using the same OLEDB src I have to lookup that table (tableB) and associate in pid with each row.
> but as the PIDs are generated by a stored procedure and are placed in a separate table, I should have some process to join with that table, but there is no relation so how can I join and get PIDs from that table to this table.
> or may be ask my stored procedue to instead of inserting it in separate table , use them in the flow return them.
any help?
Thanks
------------------
Why not ?
July 15, 2010 at 6:34 am
OK, I understand what you mean.
I'd be tempted to use ranking functions to assign a row number (order by ID, or whatever you want) to each row in both tables, then inner join them on row number - you can do the whole thing in T-SQL.
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
July 19, 2010 at 8:50 am
Hi,
If your SP does not do any DML operations, write a function instead and use it in the inline query.
Srinivas
July 20, 2010 at 5:45 am
[font="Comic Sans MS"]Instead of using a stored proc to create IDs, in the OLD DB source task, where you are selecting your records, you can use the "newid()" function as a column. This way you can do away with the other Source and merging the two inputs.[/font]
July 21, 2010 at 6:42 pm
Unless I am missing something, I think what you said below would be your best bet... just return the IDs in the flow... not sure why they need to be stored in a different table and then rejoined later in the flow.
vineet_ (7/15/2010)
> or may be ask my stored procedue to instead of inserting it in separate table , use them in the flow return them.
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply