Update analysis because it refers to items that are either no longer present

[pufaika]

Hello to all,
I have been wondering and searching for the most part of today to find an answer to "Why does after relinking package to a new package and opening the analysis studio report I get prompted to replace a missing item = measure.  Although, the new package is exactly the same.  All the dimensions and measures are the same and callled the same as well. "  If i drag the same measure in the "new item" field and then click OK to update, the report is updated but blank, there is nothing left of the report.  So what am i doing wrong or why is this happening?  Will be grateful if someone can shed some light on this issue. Thanks!

[MFGF]

Hi,

In Analysis Studio, you see the caption values of each member displayed in the package tree, but under the covers, each member is identified by its MUN (Member Unique Name).  I'd suggest that although the new package may look the same on the surface, the MUNs are different, meaning that the members used in the analysis cannot be located.

MF.

[pufaika]

Hi, ok, so how can i go about it in order to make the MUN the same?  Is that something i can do within Transformer?  Thanks

[MFGF]

Hi,

How are you saving your Transformer models - as PYJ or MDL filetypes?  If it's the latter, then the categories are not saved in the model, and get regenerated when you build your cubes.  If you have non-unique category codes in your dimensions, Transformer will make them unique by appending a ~ and a randomly generated number - this is then used in the MUNs as it forms the unique identifier for each category (member).

The way to avoid this is to save your models as PYJ - the categories are then stored in the model and are static (unless you rebuild a dimension or do a 'Clean House'.

Regards,

MF.

[prikala]

Hi,

How are you saving your Transformer models - as PYJ or MDL filetypes?  If it's the latter, then the categories are not saved in the model, and get regenerated when you build your cubes.  If you have non-unique category codes in your dimensions, Transformer will make them unique by appending a ~ and a randomly generated number - this is then used in the MUNs as it forms the unique identifier for each category (member).

The way to avoid this is to save your models as PYJ - the categories are then stored in the model and are static (unless you rebuild a dimension or do a 'Clean House'.

Regards,

MF.

A little offtopic but categories are saved in MDL files also and saving as PYJ will not make categories static (but checking "Prohibit automatic creation of new categories" will).
Best solution (imho) is to associate explicit category codes for each level and (at the data source) make sure values are unique (for example prefix them with level id, "L1", "L2" etc).

I think changing cube name (or at least package name) will also cause similar issues (invalidates all MUNs regardless of category codes etc).

[pufaika]

Thank you both for the tips.
I do store the transformer file as a PYI file.  Tested with selecting "prohibiting automatic creation of new categories"in all dimensions, but after relinking the report still got the same issue.  Although the two cubes are identical, only the package name differs  :-[

[MFGF]

only the package name differs

... and there is most likely your problem.  Is the package published solely based on the cube, or is the cube presented within a namespace within the package?  If it's the former, the package name will affect the MUNs.

MF.