Effective approach to copy data from one version to another

Navan Mehra

When we need to copy data from one version to another, we usually adopt the temp cube approach. Copy data into temp cube without the version entity and then write it to the original cube into another version. This could be quite a bit of a task if you are required to copy data for all models, there could be a large number of cubes.
Is there any other approach to do such a copy? What are others doing to if you are to pre-seed a version from another version to do scenario modelling?

------------------------------
Navan Mehra
Senior Consultant
Board Community
Australia
------------------------------

Helmut Heimann

Hi @Navan Mehra,

that's the usual approach to creating a new version.
What do you mean by "a large number of cubes"? You have one cube holding the versioned data and one temporary cube in the procedure to represent the data without the version.

Kind regards
Helmut

------------------------------
Helmut Heimann
Senior Consultant
Board Community
Germany
------------------------------
-------------------------------------------

Navan Mehra

I am referring to the number of cubes you would create to cater to different models. For example, if I have a planning model which has driver-based P&L, Workforce, and Capex, I might end up creating 50-60 cubes depending on different measures I would like to capture. If all of these cubes have a version entity and I like to pre-seed one version from another, I would end up creating 1 temp cube for each followed by a minimum of 2 steps, one to copy source data to temp and then temp to target.

------------------------------
Navan Mehra
Senior Consultant
Board Community
Australia
------------------------------
-------------------------------------------

Helmut Heimann

Hi @Navan Mehra,

yes, that might well be the case.
But, on the other hand: what would you suggest? If those cubes would only differ regarding the measure (meaning dimensioned on the same entities) you might consider using a "measure"-entity to reduce the number of needed cubes.
What do you think the system should be able to do to achieve your aim in a more efficient way?

Kind regards
Helmut

------------------------------
Helmut Heimann
Senior Consultant
Board Community
Germany
------------------------------
-------------------------------------------

Navan Mehra

I think putting them as a measure might not be possible always. Board doesn't work on a single large cube concept (like Oracle Essbase does since I have worked with that earlier). I have different measures - Text, numeric, and dates which cannot be combined in a single cube. Moreover, if such a requirement comes at a later stage of the project, it is difficult to go back to re-design everything. Anyhow, I came across another method based on hierarchies that would avoid creating temp cubes. But thought of learning from the wider group here, how they do things.

I wish there was a way to define all cubes as selection and then have a single dataflow step that would iterate through all selected cubes to perform the copy action. That would reduce the number of steps immensely.

------------------------------
Navan Mehra
Senior Consultant
Board Community
Australia
------------------------------
-------------------------------------------

Nitish Subramanian

Hi Naveen,

Using the hierarchy option sometimes impacts the processing time, in the case of cubes that are heavy in size. This experience is from Board 10.7, due to which I took the temp cube option.

Let me know how the hierarchy method works for you.

------------------------------
Nitish Subramanian
Senior Consultant
Board Community
India
------------------------------
-------------------------------------------