Answer / shrikant

Hello,
If I understand this question properly, it says that the
no. of "Source" columns are changing.
I do not agree with this scenario. Probably in
Datawarehousing, you won't find such a design. As far as
DWH is concerned, it takes the data from the OLTP systems &
after performing some operations (E-extract, T- transform)
it finally loads the data in some targets.
Here, as per the question, the question itself arises for
the OLTP design. No any OLTP (or database design principal)
system suggests a varying number of columns.
So, please do not get confused by such trivial kind of
questions. DWH is a very disciplined subject & it follows a
very good standards. Please go through the concepts first.
You will get a clear picture of DWH then.

Answer / santosh sinha

This is really a confusing question and raise question on
OLTP design butIf we are not changing the design of column
which is used in our mapping then it will not effect the
mapping so increasing the no of column will not effect.

Answer / sam

If your Source is XML this kind of scenario will come.Take
the exact XML DDL (max number of columns)as a source
defination.

Answer / suganthi

We do have a design like this(in DWH to DM load). But to
accomplish this we have created a procedure to add new
columns in teh table.

Answer / bidhar

If this is the requirement,then you will have to reimport
the source metadata in source analyzer and the save so that
the new columns are reflected in the mapping source.But to
pull it to target you need to join manually the pipelines.

But Such type of scenarios rarelly comes in OLAP models.A
good DWH design should be in such a way that if in future
any new columns come in the OLTP systems then DWH can
accomodate that.
For example Product Table has 20 columns in OLTP system.
Then OLAP model can have more than 20 columns for Product
Dimension(ATRIBUTE_1,ATRIBUTE_2 and so on)and populate null
values into that.
Now whenever new columns are added into the OLTP Product
table then we can simply map the new column to the ATRIBUTE
Columns,Keeping the datatypes in mind.

Answer / arnab

This should not happen when executing a datawarehousing
project , This reflects poor design and poor visualisation
of the requirements

The safest way to handle this is re-import the source
definition, someone suggested create a source with the
maxinmum number of columns , there is something called
quality of code and if someone codes like this it is really
a bad example of informatica coding

Answer / srinu

Some one is not agree with Sathish Rajaiah
answer(1 No)ABOVE answer.
Let me Explain Why ?

Answer / rekha

I THINK U CAN DO IT BY MAPPING PARAMETERS

Answer / sathish rajaiah

Create source defination with maximum (depends on the max
numbers of columns u get in the source) number of fields
and use that in the mapping.
ex: create source defiantion with 20 fields and use this in
mapping this wil handle source comes with columns btween 1
to 20.

Note : Above wrks only for flat file sourcse not fr RDBMS
sources

Can any one give me a real time example for FACT TABLE & DIMENSIONAL TABLE?

2 Answers  

Differentiate between sessions and batches?

0 Answers  

Plz can anybody send Informatica and Cognos FAQ Plz To thid ID ppradeep_dwh@ayhoo.co.in

2 Answers  

What is pmcmd command?

0 Answers  

error reading data in flatfiles. Instead rading Ravi Teja its reading Ravi?Teja.How to handle this issue.

4 Answers  

what is the function of 'F10' informatica ?

4 Answers   IBM,

why we are using surogate key in real time give me explanation

2 Answers   TCS,

What are the types of lookup caches?

10 Answers   Accenture, IBM, Informatica,

how to join the 2 different table with different columns in informatica?

5 Answers   Accenture,

Which tool yoU use to create and manage sessions and batches and to monitor and stop the informaticaserver?

4 Answers  

Can we make worklet inside worklet?

1 Answers  

What is the difference between view and materialised view?

4 Answers