Answer / kiran

Hai This is Kiran...

If u want to join more than one table ,u can use join,lookup
and merge also.

Join: it is used join more than one table based one key
column .it can perform 4 join as inner join,left outer
join,right join and full outer join.
Lokk-UP:it is used join more than one table,but not necesary
to join based ont he key column but it need data-type.it
give reference link and single out put link and give reject
link also.it can perform inner join and left outer jojn.

main difference: if the huge amount of the data contain in
reference table refer to join else look-up.

Answer / krishna

Generally we are using lookup for comparision purpose,
based on the reference table size we r using join or lookup.

If the reference table size is less than the main table
then use lookup stage other wise use join stage.
In lookup we have two types:
Normal lookup - reference lookup is less no of rows
compared to main table
sparse lookup - reference lookup is more no of rows
compared to main table.
But the peformance wise use join if the reference table has
more data.

Regards,
Krishna

Answer / zulfi123786

Hi This is Zulfi

Basically Join is used when you have large amount of data
about in millions and it performs inner join,left
outer,right outer and full outer joins
The join stage requires the incomming data to be hash
partitioned and sorted on the joining keys

The look up is used when the reference records are fewer in
number about less than one lakh and it doesnot require the
incomming source data to be sorted, instead the refrence
link should be in Entire partition mode.

In look up there are two types
Normal and Sparse
Sparse is available only when the reference is a database.
usually Normal has to be used unless when the refrence to
source rows ratio is 100:1

Answer / sadanand

HI All,

I would like to add one more point to JOIN.
To achieve full outer join the number of inputs need would
be only two.
The Primay table need to be sorted.
Memory used is very less compared to Lookup.

Regards,
Sadanand.

Answer / jason

that 's cool ,thank you for your sharing

Answer / indian

Hi Zulfi..you are answer is more explained one and clear

we will go for Merge if we want the rejected data for every
update link

Answer / anburadh

one more point to add on MERGE and LOOKUP
Merge can not allow duplicate records whereas LOOKUP will allow or can control duplicates using OPTION"MULTIPLE ROWS ON THE LINK" which is in constraint TAB.

How to reverse the string using unix?

0 Answers   CTS,

how to connect source to db?generally what r stages u used? how to find the data is having delimiter format?

0 Answers   Wipro,

Hi guys, 7. how to get this output i/p col1,col2 101,a 102,b 103,c 104,d o/p col1,col2 101,d 102,c 103,b 104,a Thanks

2 Answers   Cap Gemini, IBM,

What is diff between Junk dimensions and conform dimension?

4 Answers   Polaris,

job locking methods? How can we unlock the job?

3 Answers   IBM,

how to create document in datastage?

1 Answers   Cap Gemini,

Differentiate between Join, Merge and Lookup stage?

0 Answers  

How many areas for files does datastage have?

0 Answers  

Hi friends If anybody completed datastage 8.5 certification.Could please guide me what are the topics new in 8.5 certification.

2 Answers  

i 10 jobs first two jobs are runing in 2nodes,next 2 jobs are running in 4 nodes, next 4 jobs are running in 6 nodes and the remaining jobs are running on 10 nodes. how to change the node configuration?

2 Answers  

I/p : F1 table have A,B,C,D,E and F2 table having C,E,V i need output 0utput1: ABD output2: V any one suggest me this

0 Answers  

What is Horizontal transformation, vertical transformation,diagonal transformation?

2 Answers   Cap Gemini,