First of all let me point out that you can do a lot of analysis and ad-hoc, dynamic next-gen reporting with QlikView without building any Data Warehouse. But if you want to have multiple QlikView applications built for multiple areas within an organization and cross-functional and process-oriented analysis you will have a huge benefit in thinking in Data Warehousing terms and use the best of QlikView technology with best practices from Data Warehousing practices. As an example cubes are not in that category.
Q: Is it true that in QlikView we don't really need to build a datawarehouse or a datamart?
A: Yes and NO: quick answer: you don't have to but you will often end up do the some or most of the practices anyway.
You can like with a lot of other tools ignore all data warehouse disciplines and practices and just build ad-hoc.
Most people will end up with a mess. QlikView can seem magical since you can achieve a lot in a very short
amount of time. But most real-world challenges in the BI-space have messy data and multiple sources that are
not in harmony. To be able to build the right solution with the right effort - good data quality practices needs to
be followed. In the Qlik-World they don't often refer to this as data warehousing but it really is.
Q: And when is it beneficial and important to use Datawarehouse in Qlikview and when it is not needed?
A: The more complex your analysis and reporting demands are and the more unclean and dirty your source data are, the more you need a Data Warehouse - or a Data Warehouse Discipline (but you can do it all in QlikView if you like ... or combine QlikView with other tools). The collection of QVD-layers you build together with QVW-data mart models can definitly be viewed as a complete Data Warehouse solution.
Q: If we dont use datamart or datawarehouse how to historicize data in this case?
A: You have a full documentation on how to handle both history and SCD (Slowly Changing Dimensions). But not a single reference manual. The practices aren't really built into the toolset you get with QlikView. By the way Stephen Redmonds "Mastering QlikView" is an excellent book covering a lot of these practices.
Q: And last question compared to other tools, is Qlikview the only one who you can build a BI project without the concept of datawarehouse?
A: It is an illusion to think that you can do more than relatively trivial massive data lifting, data wrangling and complex analysis and reporting without employing at least a minimal set of Data Warehousing practices. You might not refer to them as such but you need to do a certain amount of them anyway.
