Dear all

I'm trying to figure out how the slow changing dimension wizard works. I
want to use it for a Datawarehouse for a Date Dimension. All attributes are
changing attributes type 1.

I expected to work it this way:
1) New rows: They are inserted with the new row task
2) Updated rows: Rows get updated
3) Unchanged rows: If the input in the excel is the same as in the db, then
I expect that the SCD Wizard sends it this way.

Unfortunately my unchanged rows always get updated.

Can anybody point me to the right direction, how to get this right?


Best regards,

Marc

My problem is, that I'm expecting, that only changed rows should be updated.
But it looks that the SCD Wizard sends all rows to the Update flow, if a
attribute is set to changing attribute. Is this correct?



Best regards,

Marc


"Marc" wrote:

Hello Marc,

Based on my scope, the unchanged rows shall not send to update flow. It
seems that you did not select the proper business key in SCD wizard.

Unchanged rows shall be added into unchanged output though there is no
default output is added by wizard.

Regards,

Peter Yang
MCSE2000/2003, MCSA, MCDBA
Microsoft Online Partner Support

When responding to posts, please "Reply to Group" via your newsreader so
that others may learn and benefit from your issue.

================================================== ===


This posting is provided "AS IS" with no warranties, and confers no rights.

Hello Peter

mmm - well it looks that the unchanged rows go through the update flow in my
package. This is what I do:
1) Extracting data out of excel
2) Get rid of null rows with a conditional split
3) Multicast to save extracted rows to raw file
4) On the other output flow of Multicast I make a data conversion of the
Business Key and the other attributes, because from excel numbers are doubles
and strings are in unicode
5) Derived Column: Add some housekeepting columns
6) SCD Component with a Update, Insert and a nonchanged flow

In my package, all unchanged rows flow to the update flow and not throug the
unchanged flow. Strange - I checked the businesskey: Its not changing inside
excel and a data conversion from double to a signed integer should not be a
problem, shouldn't it?

Well, I helped myself with a Script Component and computed a Hash Column.
This is followed by a Lookup Column, which checks if this value already exist
in the destination table. If yes, I know this row is unchanged, if the lookup
fails, I know the row is eather changed or new. This is followed by the SCD
Component.

After testing It looks to me that this path works but I guess this is max a
workaround. What do you think?

Regards, Marc


Regards,

Marc


""privatenews"" wrote:

Hello Marc,

I doubt different data type might be a problem. I copy the part of dest dim
table(lookup) as the source to another database, and choose the business
key (identity) in the SCD. No rows go to update flow under my test.

Regards,

Peter Yang
MCSE2000/2003, MCSA, MCDBA
Microsoft Online Partner Support

When responding to posts, please "Reply to Group" via your newsreader so
that others may learn and benefit from your issue.

================================================== ===


This posting is provided "AS IS" with no warranties, and confers no rights.

Dear Peter

Thanks for helping. I'm just used the SCD Wizard with a OLE DB Source and it
worked as you discribed. But with Excel Datasource, I had a problem.


Regards,


Marc


""privatenews"" wrote:

Hello Marc,

It seems Excel or type conversion shall be the cause of the problem. You
may want to test to narrow down the issue.

Best Regards,

Peter Yang
MCSE2000/2003, MCSA, MCDBA
Microsoft Online Partner Support

When responding to posts, please "Reply to Group" via your newsreader so
that others may learn and benefit from your issue.

================================================== ===



This posting is provided "AS IS" with no warranties, and confers no rights.

Peter,

well, this issue is not that important for me because I have a workaround. I
was just curious why this happend.

Best regards,

Marc


""privatenews"" wrote:

I've got the exact same problem however my data is in SQL Server 2000 not in SQL. However the fields I am transferring from are chars and varchars and the fileds I am transferrring too are nvarchar and char.

Has anyone come across this? What do I need to do?

From http://www.developmentnow.com/g/103_...it-up-rows.htm

Posted via DevelopmentNow.com Groups
http://www.developmentnow.com

Hello Katrinanospam (AT) developmentnow (DOT) com,

The SCD transform splits rows based really on two things

1. What you identify as the business key and
2. What you identify as the change type of the columns.

Data Type mismatches can be a problem in some components (try doing lookups
on floats)

Have you tried casting your columns to the same datatype as their respective
partner and see what happens?

Are you able to give a very very simple guaranteed to fail example (Recreation
scripts etc)?

--

Allan Mitchell
http://wiki.sqlis.com | http://www.sqlis.com | http://www.sqldts.com |
http://www.konesans.com