Answer / dada

Hi Guys,

We can use union transformation to join 100 tables.but the
thing is it should be same structure.by using one union we
can do it. for ex all 100 tables have same order,datatype
(std_name,loc,dept) then only we can do it by union.

Answer / bsgsr

after reading the above answers i find one is left joining
the tables in the database itself rather than joinng them
in the session using 99 joiners. it would also increase
performance.

Answer / prashant

Hello Bsgsr,

thats is one of poosible sol, we can do it only when all
the source's are Relational.
what in case if the source is a mix of Homegeneous and
Hetrogeneous ???

Answer / bidhar

If the structure of 100 tables and the target table are
same then what is the use of joiner??

Here you need to do a Union.


If the structure of 100 tables and the target table are not
same then joiner should be used.
That means you will be picking up the desired Attributes
(columns) from these 100 tables and populating into the
target.

Please let me know your question clearly.

Answer / pushparao

joiner transformation is used for combining the columns of the
different tables with common column horizontally with respect
to data of the columns.
union transformation is used to combined data for the common
columns.
or
simply joiner transformation-->Horizontal collection of the
columns.
union transformation-->vertical collection of rows.

Answer / pentaho

v can use user define join ........

Answer / sanjay

well if u hav all related sources then u can probably go
for joiner but make sure that use less no. of joiner as it
may decrease your performance.
if u hav sources with same structure and data types then u
can probably go for set operation and matchin it with your
target.
the third one is u hav to do a lot of transformation based
on some analysis and might be u can create differnet
aliases for your target and you can apply your
transformation.


plz let me know if any clarification . your comment will be
appreciatable. :)

Answer / tauseef

hai Taueef

As per above requirement my suggestion is
if the sources having similar columns then we go with union

if there is no similar columns add dummy port on all d source and use joiner transformation

Answer / kumar

Hi Gurus,

can anyone please explain performance tuning above logic. If multiple sources are coming from multiple databases.
Not simple use Unioni/ Joinder T/R.

Please be patience.. your explanation is very helpful to us.

thanks in advance

Answer / nitin tomer

Best solution as per my knowledge is below, i am assuming all tables are in DB.

1)Create a meta data table in your target DB with the three columns source_table_name,Target_table_name and Flag.
2)Use SQL transformation in informatica in Query Mode and give the required connections
3)Configure the SQL transformation and give the generic queries
create table ~Target_table_name as select * from ~source_table_name;
~source_table_name and ~Target_table_name: It will take from the meta data table

In this way you can automate this process and can load as many tables you want just you have to give table name entry in meta data table.
Write me @thisistomer@gmail.com if you have any doubts.

What is diff between connected and unconnected lookup?

5 Answers   BirlaSoft, Blue Star Infotech,

Define informatica repository?

2 Answers   Wipro,

explain abt ur project architecture?

0 Answers   Anthem, DELL, IBM, TCS, Wipro,

How informatica process data means how informatica works?

2 Answers   TCS,

Please let me know how to make encryption and decryption with example?

0 Answers  

what is dynamic cashe

3 Answers   Informatica,

what are the reusable tasks in informatica ?

11 Answers   CTS, HP, IBM, TCS,

The structure of source file is as below: Source structure(two fields) Name, Card NUmber A, 111111111(SSN) A, 01010101(Creditcard number) A, 34343434(Debit card number) B, 55555555(Creditcard number) C, 77777777(Debit card number) Target Structure(4 fields) Name,Credit card,SSN,Debit card A,01010101,111111111, 34343434 B,55555555,, C,,,77777777 Corresponding to one name there can be maximum 3 rows and minimum zero rows. Given that I do not know which record might have a particular type of number. How can I handle above requirement with informatica transformations?

1 Answers   Amdocs,

in performance wise union is past,union all is past

2 Answers  

Why do we need informatica?

0 Answers  

What do mean by local and global repository?

0 Answers  

what is parallel querying and what r hints.

1 Answers