Can you please help me out ..... 1

[ProDev]

Hi,

Written below are the few situations that I sometimes come across during my developments.Can you pls help me out on this ......


Q1.)

Here is some thing that I want to do in Informatica.

After a load is done to my target table, I need to insert
a last row, which might be formed at some expressin
transformation and brought to target.

How to do this ? Do I have to have separate mapping
to be run after my 1st map ?? OR in a single mapping
I should be able to handle this thing ?


Q2.)


I am runing 2 sessions.
1st one does populate target table T1.
2nd one does the same, once 1st is done with.

I would like the Seq No value need to be continued
from the one that is left by 1st mapping.

I mean to say that, my 1st mapping is using Sequence Generator transformation to load the respective field "SeqNO" at T1.

So, how would I be able to continue it in my 2nd mapping ? (if I use a Sequence Generator transformation in mapping 2, I get different nos, not the continuatuion.)


Q3.)

How can I know the no of rows that pass through certain
transformation, inside the mapping ? Let me know if you have got any work-around ro so.


Q4.)

How do you promote your mappings to QA environment ?
Do you just mantain a separate folder having required
rights for QA guys and copy your development mappings one by one or the whole folder ?


OR, a separate repository in mantained for QA and you
copy your development folder to QA repository ? (That cass
how to distinguish between mappings that are from developments ?)

So, what is the real procedure that you guys follow up at
your work place ??



Thanks

 

[blom0344]

Q1

The whole concept of using batches is to establish a sequence in series of sessions. So writing information to a target table AFTER an insert can easily be performed with a second session (using another mapping)

Q2-Q3

You may have a perfectly good reason for wanting to do what you are asking a solution for, but the order in which data is stored within a table SHOULD be inconsequential.
With proper datastructure getting information from a table should not depend on WHERE data is stored..........

If you are anxious to know the order in which you are storing I would suggest writing a timestamp to the table.
If you combine timestamp with sequence you can still deduce the order over all the inserts.......

T. Blom
Information analyst
tbl@shimano-eu.com

 

[Sridharan]

ProDev,

Regarding Q2 answer make a Sequence Transformation as a Reusable Transformation and set the no. of cached values to 1. Then use this in the two mappings and the second session will continue the no. from the end of the first run mapping session.

Sri

 

[blom0344]

Sri,

Did you test whether the second session will take on from the last value of the first?
It sounds unlikely to me cause each sequence generator has to be set within each mapping....

T. Blom
Information analyst
tbl@shimano-eu.com

 

[Sridharan]

blom,

It will provided one makes the Seq. Gen Transformation a Reusable one and use the same thing in both the mappings. Also one has to set the no. of cached values to 1 to avoid the second session to use the values outside the cached values.

I did a quick testing and it works fine. Maybe you can also do a test and confirm this.

Regards,
Sri

 

[blom0344]

Sri,

I ran a test with two mappings.
Learned something (and that's where forum's are all about)

Here is a star............

T. Blom
Information analyst
tbl@shimano-eu.com

 

[Sridharan]

blom,

blom wrote:
"Learned something (and that's where forum's are all about)"


Yes definitely.

Also nice to have received a star from you.

Regards,
Sri