TM1 Forum

Hi,
Can any of us highlight the use of adding new dimension in existing cube, when after adding dimension we have to re-write all the rules and re-build the TI process.
Which takes the same time and efforts as we create new cube.

Is there any use of this functionality ??

Thanks

Tm1andTm1 wrote: Is there any use of this functionality ??Thanks

In sales demos yes! Not sure about real life development.

How it is useful in sales Demo.....???

It's a step in the direction toward useful. One would assume down the track it'll rewrite all the rules for you.

Lets hope for the best, it could re-write all the rules after adding or deleting the dimension from the cube..

Tm1andTm1 wrote:Hi,
Can any of us highlight the use of adding new dimension in existing cube, when after adding dimension we have to re-write all the rules and re-build the TI process.
Which takes the same time and efforts as we create new cube.

Is there any use of this functionality ??

Thanks

In a word - none. It's sales bling to woo mangement in to parting with cash, something which has been a trend recently. It might save a bit of time in some aspects, but its a task that is done so infrequently in a real life production system that i find it a waste of their development time when more useful things could be fixed...

In the event you have only just built the cube (with no rules or data etc) and made a mistake it could come in handy... however it takes all of about 5 seconds to delete the cube and then build again from scratch so.......

Thanks Declan,

I was wondering how it can write/update rule if you delete or add dimension in cube.......

I think it was useful in EP where rules are written at dimension level. Even deleting or adding dimension in cube does not affect to the rules because it is written at dimension level.

Good step towards EP functionality, hope this will continue adding EP functionality in TM1. Thanks to IBM

Steve Vincent wrote:In a word - none. It's sales bling to woo mangement in to parting with cash, something which has been a trend recently. It might save a bit of time in some aspects, but its a task that is done so infrequently in a real life production system that i find it a waste of their development time when more useful things could be fixed...

[off topic]
TM1 Web for example, in version 10.2 it's a disaster. Nothing more, nothing less. Oh yes, a nightmare, that as well.
[/off-topic]

Hi Wim,

Can you explain what is disastrous in TM1 10.2 web.

Hello

Making an Active form with 1 dimension for the rows (which has an MDX subset that is selected to be used in the Active form).

Then I added some formatting (dark blue, light blue, ...) in the format range for the active form.
Also changed some basic formatting in that range, like decimals, center horizontally, percentage formatting, and so on.

In Excel it works fine, well, most of the time. Problems start when level 0 cells do not accept input (that is, the DBRW shows 0 again after inputting a value and hitting Enter).
All the below was observed in TM1 Web:
String cells become numeric cells.
The title (admittedly, it's 4 cells merged together) shows up wrong most of the time.
The blue shades "change places".
Freeze "pain"s ...
A value entered at leaf level consolidates in 1 direction (dimension) but not in a different one (and the cube viewer shows the correct values).
Having 2 active forms uploaded, you open one after the other: invalid operation (or something like that, don't remember). Open 1 and close and open the other, it works and no such error. Open up the second again (so 2 files open at the same time: error.

The MDX expression used is just the children (till level 0 of a hard-coded consolidated level element).
Changing the MDX subset to a static one seems to lead to less problems, but still a number of problems mentioned above, apply.

Just another weird one. You have an Active form with MDX driven subset. Then insert a number of empty rows in between the format range and the active form.
In that "open space", use DBRW formulas to retrieve data from a cube, on the same elements as you had below. For example just copy some rows in the active form to above.
Also, from the same dimension, retrieve data from data cube but use an element that was NOT part of the MDX.
Result: (only) these extra elements and their associated DBRW's will loose their (manual) formatting on the Web. Cells that contain a DBRW for these extra elements, will be empty in the Web, while the DBRW is there in Excel.

And tens of other (what appear to be) bugs / problems in this first release of TM1 Web.
From time to time it helps to clear the cache in IE while developing the websheets.
Also, Firefox seems to do a better than IE for TM1 Web.
Lastly, updating Java can also help sometimes.

Do other people find the same?

Hi

Unfortunately I had to add in an extra dimensions to an existing cube, that had already been in use for many years. The problem was not so much with the rules as with all the related items such as Views, and Spreadsheets.

The client I was working for needed an additional split in the cube. The only feasible way was to add another dimension.

I had to do the following.

a) Create the new Cube with the extra dimension on another server.
b) Write processes to export data from the old cube and load it to the new cube. They were content to just map the data from the old cube to an Unknown element in the new dimension, and then just start recording the new split once the new cube went live.
c) The rules were a relatively simple Find and Replace, to add in !<new dim>
d) I used one of my API routines to generate the TI statements needed to create all the Views on the old cube. I then used some Excel formula to add in a line to assign a subset on the new dimension to the view, with the Top level element of the new dimension selected for the Display type views. I then pasted this into a TI process and ran it to create the views on the new cube. (It would have been easier if the views had been created by TI in the first place but they weren't, and in any cube that has been around a while, they probably won't be).
e) I then wrote some VBA to scan all the existing reports for references to this cube. It then added in a line to the sheets to hold a SUBNM formula with the Top level element of the new dimension, and it updated all the DBRW etc references to reference the additional element.
e) Finally I needed to change the load process but there were only two processes involved so like the rules that was relatively easy.

I can see a day when Performance Modeler might do the rules, views and processes. However, I doubt that it will be able to do the spreadsheets which was the hardest part.

Regards

Paul Simon

Thanks So Much Wim for valuable information...

Wim Gielis wrote:Hello

Making an Active form with 1 dimension for the rows (which has an MDX subset that is selected to be used in the Active form).

Then I added some formatting (dark blue, light blue, ...) in the format range for the active form.
Also changed some basic formatting in that range, like decimals, center horizontally, percentage formatting, and so on.

In Excel it works fine, well, most of the time. Problems start when level 0 cells do not accept input (that is, the DBRW shows 0 again after inputting a value and hitting Enter).
All the below was observed in TM1 Web:
String cells become numeric cells.
The title (admittedly, it's 4 cells merged together) shows up wrong most of the time.
The blue shades "change places".
Freeze "pain"s ...
A value entered at leaf level consolidates in 1 direction (dimension) but not in a different one (and the cube viewer shows the correct values).
Having 2 active forms uploaded, you open one after the other: invalid operation (or something like that, don't remember). Open 1 and close and open the other, it works and no such error. Open up the second again (so 2 files open at the same time: error.

The MDX expression used is just the children (till level 0 of a hard-coded consolidated level element).
Changing the MDX subset to a static one seems to lead to less problems, but still a number of problems mentioned above, apply.

Just another weird one. You have an Active form with MDX driven subset. Then insert a number of empty rows in between the format range and the active form.
In that "open space", use DBRW formulas to retrieve data from a cube, on the same elements as you had below. For example just copy some rows in the active form to above.
Also, from the same dimension, retrieve data from data cube but use an element that was NOT part of the MDX.
Result: (only) these extra elements and their associated DBRW's will loose their (manual) formatting on the Web. Cells that contain a DBRW for these extra elements, will be empty in the Web, while the DBRW is there in Excel.

And tens of other (what appear to be) bugs / problems in this first release of TM1 Web.
From time to time it helps to clear the cache in IE while developing the websheets.
Also, Firefox seems to do a better than IE for TM1 Web.
Lastly, updating Java can also help sometimes.

Do other people find the same?

Thanks Paul for valuable information, So we need to create new cube with the new dimension we want to add.
There is nothing we can do using performance modeler. There might be some other use of this functionality because IBM might have evaluated the functionality before implementing or adding it to TM1. Which we are not able to understand now.

paulsimon wrote:Hi

Unfortunately I had to add in an extra dimensions to an existing cube, that had already been in use for many years. The problem was not so much with the rules as with all the related items such as Views, and Spreadsheets.

The client I was working for needed an additional split in the cube. The only feasible way was to add another dimension.

I had to do the following.

a) Create the new Cube with the extra dimension on another server.
b) Write processes to export data from the old cube and load it to the new cube. They were content to just map the data from the old cube to an Unknown element in the new dimension, and then just start recording the new split once the new cube went live.
c) The rules were a relatively simple Find and Replace, to add in !<new dim>
d) I used one of my API routines to generate the TI statements needed to create all the Views on the old cube. I then used some Excel formula to add in a line to assign a subset on the new dimension to the view, with the Top level element of the new dimension selected for the Display type views. I then pasted this into a TI process and ran it to create the views on the new cube. (It would have been easier if the views had been created by TI in the first place but they weren't, and in any cube that has been around a while, they probably won't be).
e) I then wrote some VBA to scan all the existing reports for references to this cube. It then added in a line to the sheets to hold a SUBNM formula with the Top level element of the new dimension, and it updated all the DBRW etc references to reference the additional element.
e) Finally I needed to change the load process but there were only two processes involved so like the rules that was relatively easy.

I can see a day when Performance Modeler might do the rules, views and processes. However, I doubt that it will be able to do the spreadsheets which was the hardest part.

Regards

Paul Simon