If you are unable to create a new account, please email support@bspsoftware.com

 

How templates and lookups are related?

Started by larus, 18 Sep 2012 04:09:03 AM

Previous topic - Next topic

larus

How templates are related to lookups in the dimension reference structure. How they are used in Fact and Dimenion builds? Should I create a template before I create a lookup for a table? Are lookups only used in Dimension builds? When to use template acccess in lookup and when to use datastream access?

MFGF

A template is simply a list of attribute names. They are used with dimensional structures - hierarchies and lookups - to define the names of the attributes used in those structures. When creating a lookup, you *must* have a template defined which provides the attribute names. These can be the same as the names in the underlying data source(s) or they can be different - your choice. When creating a hierarchy you can have one template per level, or one for the whole hierarchy, or any mix between those two extremes. Again, the template(s) describes the attributes used on the levels, and the same rules apply - the names can either match the names defined in the underlying data source(s) or not; it's your choice.

Template access for lookups and hierarchies is a fast, simple method of getting Data Manager to construct your source query behind the scenes without you having to define it. There are some pretty restrictive conditions for when you would use Template access, though. It can be used only when:
1. All the data to be read is in a single table
2. The names of the attributes you need to read (in the template) match the column names in the table exactly.
3. You do not require any derivations/pivoting/weird mapping to be done as the data is being read.

Otherwise, you will need to define data source access and create a data query, map the columns from this to the datastream then map these items to the attributes in your lookup/hierarchy level.

Templates are also used in dimension builds. Here, the attribute names in the template are used to describe the column names in the dimension table. Each attribute can also have a behaviour defined, and this behaviour controls how Data Manager treats the relevant column in the dimension table (surrogate key, business key, effective date etc)

Templates are not used directly by fact builds. They are used indirectly in any lookups or hierarchies the fact build references, and in any dimension deliveries it may make, but that's all.

Cheers!

MF.


Snet form my fumblefingers
iPhon 5 usig Tapatalk
Meep!

larus

Many thanks for this profound answer!