if we take 2 tables(like emp and dept), we use join stage and
how to improve the performance?
Answers were Sorted based on User's Feedback
Answer / kiran
when ever join 2 tables based on key columns if the key
column is numeric ,set modulus,if the key column is non
numeric set hash partition technique.and compare to look up
join give better performance coz join has sort operation
by default.
| Is This Answer Correct ? | 11 Yes | 2 No |
Answer / ashok
above answer has one mistake
i.e join doesn't has sort operation bydefault we explicitly
specify
| Is This Answer Correct ? | 9 Yes | 3 No |
Hi this is Poorna ,
We can Improve the performance on join stage by doing
pre sorting for both left and right data based on
key .Then we can Improve the performance in join stage .
Plz correct me if any mistake in thinking .
| Is This Answer Correct ? | 6 Yes | 1 No |
Answer / rajeshchunduri
in emp and dept tables key column is deptno so it is key
based and datatype for key column is int . At this time we
change partion tech from hash to modulus.
chunduri
| Is This Answer Correct ? | 1 Yes | 1 No |
Answer / professional
Hi,
For the above query to improve the performance based on key columns in emp and dept joins by default sort in datastage for better performance if you have already a sorted data just go for environmental variables and do the operation #APT_Not_SORTDATA option then performance increase automatically...
| Is This Answer Correct ? | 0 Yes | 0 No |
how many write modes are there in ds
Have you used Unstructured data?
How can we do null handling in sequential files
Can we use target hash file as a lookup ?
What are the different types of lookups in datastage?
what is the Difference Between Datastage Server Edition and Parallel Edition?
What all are the different way to run a job?
In merge stage can we update the values?
Difference between sequential file and data set?
How do you import and export data into datastage?
what are .ctl(control files) files ? how the dataset stage have better performance by this files?
What is difference between symmetric multiprocessing and massive parallel processing?