1. What is stiched Query
2. What is loop in Fm
3. Best practices in modeling in FM
4. Three types of SQL
5. Difference between pass-thro and native sql?
6. will usage of pass thro sql will improve the performance?
7. Complex problems fced in your project
8. performance tuning in reports



1. What is stiched Query 2. What is loop in Fm 3. Best practices in modeling in FM 4. Three types..

Answer / bob

1. What is stiched Query

Stitch query is a way in cognos to avoid full outer joins.
If you are trying to model a full outer join involving two
fact tables related through a common dimension table, note
that Cognos 8 will handle such situations using a stitch
query, which does not require that you model any of the
relationships as an outer join.

If you have a forecast for a new backpack called the Day
Hiker 3, you may not have any sales yet for that product.
You might have some sales for an older backpack, say the
Day Hiker 1, though you might not have forecast any sales
for it because it is no longer being promoted. The Day
Hiker 2 backpack might have both sales and forecast
information. If you create a report for products, sales,
and forecast, and if you have properly modeled the
situation, the report will produce a stitch query. That
stitch query will first retrieve the sales data for the Day
Hiker 1 and the Day Hiker 2 backpacks, as well as the
forecast information for the Day Hiker 2 and Day Hiker 3
backpacks. Stitching those two results sets together,
Cognos 8 will display that you have only sales for the Day
Hiker 1, sales and forecast for the Day Hiker 2, and only
forecast for the Day Hiker 3. This is effectively a full
outer join without having to model explicit outer joins in
the relationships among these tables. You will notice a
coalesce statement if you look at the generated SQL in a
report that uses a stitch query. (Unserstand the example)

2. What is loop in FM?

A Loop is a closed path (relation) that exists among 3 (or)
more tables. For example, if we have '3' tables P1, P2, P3
then, a loop exists among these tables only when we create
joins in the following fashion:
Loop: P1 -> P2 -> P3 ---> P1 (hence, forming a loop)
To resolve the above problem, we have to create a shortcut
(or) Alias to the Table P1.
No Loop: P1 -> P2 -> P3 -> Alias (or) Shortcut of P1
(Remember. Interviewer will ask you abt how to create short
cuts or alias and hw many of them are there; Prepare for
that also:))

3. Best practices in modeling in FM?

There are many best practices that you can use when
modelling your data. And, you have to go through some of
the web sites to get this info. I have collated a few of
them and they are project specific:

•Login as Read Only user and check out the model as local
copy, once the Model is ready and developed use the
licensed user to merge your changes
•Once the Model is updated please validate your Model by
generating report
•Add Description for the new version which is created by
you. Helps in tracking the changes.
•Re-Check the default value of a column once the Model is
updated.
•Provide Latin as character set if Unicode is not
applicable.
•Provide the Owner Name for all objects, example EDW_DATA
(not development Owner Name: FIN3_ETL_TARGET)
•Specify the relationship among the objects, try not to
hang any objects.

Tell gthe interview that this were the giude lines in youer
modelling.

4. Three types of SQL

There are there types of SQL in Cognos especially working
with FM. They are:

a) Native SQL: Native SQL is something that is generated as
per the terms of the SQL language. Here, all the iformation
are retrieved when going for select statement and hence
somtimes works as similar to Select * statement.

b) Cognos SQL: This SQL is generated by Cognos with some
modifications and have improved performance over Native SQl
since, it ensures that it is only going to retrieve trhe
values that are requested by you and not all.

c) Pass Through SQL: This tyoe of SQL is not at all
modified and sent to your DB for querying and hence, should
not be used as this degrades the performance when compared
to the Cognos SQL.

8. Performance tuning in reports:

1. Don’t use the Not in function in the query.
2. Check the Governor Settings.
3. Check the table qualifying and weighting.
4. Check the client/server balancing.
5. Avoid using complex calculations avoid complex
reportcheck filter condition, use less filter, avoid
unwanted calculation

Is This Answer Correct ?    12 Yes 1 No

Post New Answer

More Cognos Interview Questions

What are the two types of framework manager?

0 Answers  


what is filter?give me one example that you created a report in your project?

1 Answers   HCL, IBM,


i have one file i saved in csv how can i take these file into framework manager

0 Answers  


How to create a dynamic column name in Cognos?

6 Answers   Aditya Birla, Polaris,


How do you create IQD in FWM? Pls, give the procedure with an example. I told him that we could create using Native SQL method in FWM; it is correct or do you do another method there?

4 Answers  






Can you define report item?

0 Answers  


Explain what is the main advantage of impromptu over cognos reportnet?

0 Answers  


What are cognos connection initial page components?

0 Answers  


we have a column of int/varchar datatype,but when we run the report we should get the date format i.e., suppose date key is a query item consists of data as 20040105,20030503,20060410....... output should be 5 jan 2004,3 may2003,10 apr 2006..........respectively.Can i knnow the answer

4 Answers   IBM,


Where do you set security levels for users in reports?

5 Answers   HCL,


what is the difference between cascading prompt and drill-thru

4 Answers  


suppose a value prompt contains 100 values and i want to select all values in that prompt....i want ALL option in the value prompt? how can u keep this option in value prompt?

5 Answers   TCS,


Categories