Question { IBM, 7381 }

can we use cartesian join in informatica

Answer

we can use Cartesian join in informatica by using full outer
option in joiner t/r


if am wrong please let me know

Question { 5984 }

what is the max/min size allocated for caches.whether index
or datacache in any of the transformations like Joiner or
Aggregator.bcoz it is set to Auto by default which means we
need not to allocate no of bytes or MB.so what is the use of
these properties.please let me know about the caches.

Thnaks in advance
Sai Ram

Answer

INDEX CACHE IS MIN IN SIZE AND DATA CACHE IS MAX IN SIZE

Question { IBM, 8995 }

which transformation should we use to get the 5th rank
member from a table in informatica?can we achieve this in
sql?

Answer

FOR THIS WE HAVE 2 USE TWO T/R ie FIRST WE HAVE 2 USE RANK T/R
AND THEN USE A FILTER T/R IN FILTER GIVE THE CONDITION AS
RANK=5 CONNECT TO TARGET
THE FLOW IS LIKE THIS
SRC --->SQ--->RANK--->FILTER--->TRG

WE CAN ALSO DO THIS IN SQL USE THE FOLLOWING QUERY

SELECT * FROM WHERE ROWNUM <= 5 MINUS SELECT *
FROM WHERE ROWNUM <=4;

Is This Answer Correct ?

8 Yes

0 No

Question { Emphasis, 8040 }

i have n number of records in my source, i want first and
last record to my target. how can u implement this scenario .

Answer

we can also do it like this
FIRST PIPELINE

From sq connect all the ports to expression t/r and by using sequence generator t/r generate sequence number.
From expression t/r connect the all the ports to the filter t/r write a condition sno=1.
from filter t/r connect all the ports to the target
SECOND PIPELINE

From sq connect all the ports to Aggrigator t/r and dont group by any port,connect the ports to the target
THE DATA FLOW DIAGRAM WILL BE LIKE THIS
sequencegenerator
|
\|/
src--->sq--->expression--->filter--->trg
|
|-->aggrigator--->tgr


I think this will work

Question { Wipro, 7820 }

What is the exact difference b/w IN and EXIST in Oracle..?

Answer

EXIST is used as conditional operator in which u can write a query
IN is also used as a conditional operator in which we gives a list of elements

Question { Satyam, 5993 }

my source having the records like

ram 3
sam 5
tom 8

and i want to load into target like ram record 3 times,sam
record 5 times, tom record 8 times

Answer

let us take the metadata as name and records ok,
in expression t/r we create a new o/p port in that write the
expression as name ||''||record||''||records||''|| times


connect the newly created o/p port to the target.

Question { Cap Gemini, 12329 }

I have a file it contain 2 records like empname,company as
Ram, Tcs and Ram, IBM. But i want empname,
company1,company2 as Ram, TCS,IBM in the target. How?

Answer

The flow will be like this

src----->sorter---->transformer----->remove duplicates---->trg
SORTER
in sorter sort by the key=empname and create keychange column

TRANSFORMER
create two stage variables stg1 and stg2
if key change=1 then company else stg2 :',': company |stg1
_____________________________________________________|_____
|
stg1 |stg2

connect the stg1 as out put column and empname as output

from transformer connect to the remove duplicates
REMOVE DUPLICATES
key=empname
duplicates to retain=last
connect to the trg

u will get the output

Question { Cap Gemini, 10957 }

source target
--------- --------
name no sal name no sal max_sal
r 10 1000 r 10 1000 3000
y 20 3000 y 20 3000 3000

Answer

source-->sq-
| \-------------|
| --->aggr--->target
in target create a new column maxsal now drag the source and
target to the mapping area connect the source to sq.from sq
create two pipelines connect the first pipeline to the target
and the second pipeline to the aggr t/r in aggr t/r create a
new port max_sal=max(sal) connect this pipe line to the target

Question { IBM, 28806 }

How to generate surrogate key without using surrogate key stage?

Answer

by using the transformer we can do it.To generate seqnum
there is a formula by using the system variables
ie
[@partation num + (@inrow num -1) * @num partation

Question { TCS, 6312 }

Source-1
No name
1 satish
2 karthik
3 swathi
4 keerthi

Source-2
No name
1 satish
2 karthik
5 santhose
6 vasu


Target


3 swathi
4 keerthi
5 santhose
6 vasu here i want non matching
Records i want how to achieve that

Answer

here as the source meta data is same we can use UNION
t/r,after that i use AGGRIGATOR t/r in that i will count(no),
after that i will keep the filter t/r in that the condition
is count=0.then connect to the target
the result is as u want
the flow is as fallows

src--->sq--->union--->agg---->filter---->trg

Question { Thomson Reuters, 12037 }

in my source i have 100 records, and 3 targets, i want to
load
1st record into 1st target,2nd record into 2nd target,3rd
record into 3rd target again 4th target into 1st target and
vice versa,how to achieve this?

Answer

here we want to conform one thing that's is there serial no
or not
if there is serial no then we can just take a router then
create three(3) groups
First group for First target
Second is for second target
Third is for third target
write the condition as fallows
mod(s.no,3)=1-------> first target
mod(s.no,3)=2--------> second target
mod(s.no,3)=0-------->third target
if there is no s.no in source then create s.no by using
sequence generator t/r


the flow will be as fallows
/--------->t1
src--->sq---->router/----------->t2
\------------>t3

or if u use sequence generator the flow will be

sequqncegenerator
| /--------->t1
src--->sq---->router/----------->t2
\------------>t3

Question { TCS, 17222 }

Please explain me the difference between 3 types of slowly
changing dimension in datawarehousing?

Answer

scd1 ---->will contain only the updated data(ie it contains
the newly entered data + updates of historic data)it dose
not maintain historic data

scd2 ---->it contains the updated data and the historic data
(full history)
scd3 ----->it contains updated data and historic data
partially(means it may contain the records of the last
6months i think)

Question { TCS, 8722 }

Can we use Round Robin for aggregator?is there any benefit
underlying?

Answer

we can't use round robin for aggrigator
Because round robin shares the records equally to all the
nodes.
but if it goes like that there may be a chance to get not
relevant o/p (UN expected o/p )

its better to use hash partition



if am wrong plz let me know