TM1 Forum

Hi guys,

I've recently tried to recreate an exact duplicate of a view extract (with the intention of simply using exactly the same source as an existing view extract but with a different name) and during the reboot process of the instance the view extract has changed to the point that it no longer represents the original source view. Here's the background:

Lets call the source view sView and the target view tView. Both of these views are of the same cube. In using the View extract window (from the datasource tab in the TI editor clicking on "Browse...") I create the same subset selections or element selections per dimension and then save this view extract with a new name (tView). When I open this view extract again it looks to be the same as the source view extract so all looks good so far.
When I view both of these from the cube viewer they look different: sView has 2 of the 14 dimensions in the title area, 1 in the column area and the rest in the row area. tView has 1 dimension in the row area, 1 in the column area and the rest in the title area.

The exact same element or subset selections were made in creating these view extracts though, so the question is why are these view different to each other when viewed in the cube viewer and what is the preferred method of creating a duplicate/clone of an existing view extract?

Thanks for your input.

You're not creating a "View Extract", you're just creating a simple view. Any view that is created manually will have whatever dimensions are in the Title area at the time the view is saved. If you click on Browse from the Ti editor window (which BTW qualifies as creating manually) then you are just using the default view for that cube, whatever it is. If that default view has dimensions as title then that's what you'll get in your saved view. To create the type of view normally seen in a TI (which you might refer to as a View Extract) then you should either create it in code, with the ViewCreate and ViewSubsetAssigns commands or manually create it by right-clicking on the cube and choosing "Export as Text Data", make your dimesions selections and then Save As.

Thanks Tom for that insight.

I assumed from reading other posts that when a TI is using a cubeview in it's datasource tab that it treats that view as a view extract. What is the difference between a view and a view extract? A google on this doesn't return much, in looking this up in the tm1_TurboIntegratorGuide.pdf (pg29 of 112) it says that "browse" button opens the "View Extract Window" which creates the view. Is a view different to a view extract and if so in what way?
thanks for the pearls!

JamiseBondi wrote:I assumed from reading other posts that when a TI is using a cubeview in it's datasource tab that it treats that view as a view extract. What is the difference between a view and a view extract?

A view created in cube viewer is intended for human consumption. It usually has titles, rows and columns, though of course it can have only two or even only one of those.

The View Extract (Export As Text Data) dialog is intended to create a data source for code to process. If you create a view extract from the outset in that dialog (that is, you create it without loading an existing view and modifying it), then all of its dimensions will be rows since obviously a data source should be in a flat format. Generating a flat data file is the only reason that you should ever use the View Extract dialog. You definitely should not use it to modify Cube Viewer views. If you do load an existing view that was created in Cube Viewer in the View Extract dialog then it will probably retain some of its past characteristics such as titles and columns, with the only semi-predictable results that you have seen.

Either of these types can be created in TI through the judicious use of the ViewTitleDimensionSet, ViewRowDimensionSet and ViewColumnDimensionSet functions.

There are many (me included) who believe that Iboglix should have distinguished between those two types of views. Given that this has been "on the radar" for over 9 years now, don't hold your breath.

Now the only question to be resolved is what earthly reason you could possibly have for needing two identical views under different names. The reuse of a single resource seems more efficient and less error prone than the duplication of such a resource. Lean methodology, and all that.

Now the only question to be resolved is what earthly reason you could possibly have for needing two identical views under different names. The reuse of a single resource seems more efficient and less error prone than the duplication of such a resource. Lean methodology, and all that.

Hi Alan,

Thanks for your reply. It does explain the use of view extracts, I had muddled up the "Export as Text Data" function with the cube viewer "View Extract" window which is to be used to modify existing cube views.
To answer your question: A change was requested that required a slight naming convention change and in keeping all associated objects in sync with each other (cube views, subset names, TI names and output files) I wanted to keep all objects using the same naming convention. The original source view will be deleted.

JamiseBondi wrote:A change was requested that required a slight naming convention change and in keeping all associated objects in sync with each other (cube views, subset names, TI names and output files) I wanted to keep all objects using the same naming convention. The original source view will be deleted.

Sounds like a case where using python or a powershell script to change file names and text within files might be an option worth exploring.