LAG function in Performance Modeler

[Mark H]

Folks, the rules generated in PM using the LAG function are not working as described in the TiPs sections. Will not accept instruction.


LAG Function in Performance Modeler:

Dimension Members:

-Starting Inventory
-Opening Inventory
-Movements
-Closing Inventory

My Time Dimension is set up with each element having a PreviousPeriod other than the first one.

'Starting Inventory' is to be used as a PAD

'Opening Inventory' = LAG('Starting Inventory', 'Closing Inventory'). [Where Starting Inventory is to be used as a PAD item]
Closing = Opening+Movements


PM will accept these combinations:

1. LAG('Closing Inventory') Here there is no PAD item
2. LAG(100, 'Closing Inventory'). Here 100 is a hardcoded PAD item

The help/tips sections states that we can use another dimension element as a PAD. But it just will not accept it.

LAG calculates a result in one row by lagging an input from another row by 1 period. For example, the member ‘Opening Balance’ may use a function =LAG(‘Prime value’, ‘Closing Balance’). The <Pad> argument specifies the value returned by LAG for the first leaf member in the Time dimension; it may be another member in the dimension or a constant. If it is omitted, the user may key a value for this function into the first leaf member of the Time dimension.

Rule Generated using Hardcoded 100 PAD:

['S_Replenishment':'Opening Inventory']=N:IF(ATTRS('S_Time Yrs', !S_Time Yrs, 'previousperiod')@='',
100, DB('Replenishment', !S_Products, !S_Approval Hierarchy, ATTRS('S_Time Yrs', !S_Time Yrs, 'previousperiod'), 'Closing Inventory'));
#EndRegion

Trying hard to use PM, not the flexible TM1 rule editor method.

Thanks

[tomok]

PM is for people who don't know how to write rules themselves. I wouldn't touch it with a ten foot pole. Why use paint by numbers when you are a successful artist?

[Mark H]

@tomok : that hurts! PM is the future of TM1. All modern day developer tools have a short-cut interface of some sort. It is the sole reason why there has been such a recent resurgence on TM1. I know two very large london based banks that waited until 10.1 before taking the TM1 plunge and of course it is generated work for all of us. The latest niche is migrating Cognos Planning (CP) onto TM1 using PM links and calcs. As you know this is effectively CP inside TM1 - the clear direction that IBM are taking.

So I prefer to be part of the future and attempt to influence the next release by raising my own change requests/enhancements. If we dont use it they will not receive any change requests. My current project gave me a strong mandate that i was to use the 'new tool' and didnt want any hard coding round the edges. Why? there arent any available TM1 Masters to look after it due to UK wide demand on TM1 resources....

If a client has the message that TM1 can be built quickly and does not require a 'TM1 Master' to look after it then they are more likely to buy it - right? Thanks to the new version and PM the TM1 mesage is changing - we are shifting from these sound bites:

"TM1 = IT owned projects, but IT dont speak 'finance' "
"TM1 requires team of highly skilled people to look after it - cannot built it and leave it"
"for TM1 you need a keyboard, but Cognos Planning you just need a mouse"
"TM1 is deployed via the dreaded excel interface and spreads like a virus through a enterprise - IT dept doesnt know whats in the organisation so can't plan for upgrades" [This is direct quote from a recent customer who has had TM1 for years.]

to these:

"TM1 can be finance owned - useful when it is finance who buy the software, esp for budget/forecast systems rather than simple consolidation builds"
"TM1 10.1 = CP inside TM1 meaning whole new tranche of projects thereby raising the profile world wide"
"Combining the CP ease of building with the power of TM1 engine is a game changer"
"TM1 seemlessly intergrating to Cognos BI - another new user community for TM1"
"TM1 can be build it and leave it"

If you were the sales guy which story would you rather tell?

We must recognise that since the introduction of TM1 Contributor this toolkit will be used more and more for Planning systems instead of more common place TM1 reporting projects. Planning = finance community and finance systems super users who need to be trained to make changes on PM rather than very long multi-dim calcs of native TM1.

.....and relax. Thanks for your replies to my posting please keep them coming.

Cheers

M

[tomok]

Why? there arent any available TM1 Masters to look after it due to UK wide demand on TM1 resources....

That's because it is quite a complex tool and there are no shortcuts to building a solid, scalable, performing, model. PM is not going to change that any time soon.

We must recognise that since the introduction of TM1 Contributor this toolkit will be used more and more for Planning systems instead of more common place TM1 reporting projects. Planning = finance community and finance systems super users who need to be trained to make changes on PM rather than very long multi-dim calcs of native TM1.

I've been using TM1 for almost 17 years now and I can honestly say that about 50% of my projects have been planning and forecasting, 20% other type of dynamic calc type models and only about 30% reporting only. I don't know where you've gotten the impression planning with TM1 is only just beginning to emerge. Planning existing on TM1 long before Contributor was introduced.

The TM1 engine is very flexible, yet complex and with it being RAM-based, it is crucial your model be as efficient as possible to perform at an acceptable level and be scalable. I could never use PM at a client with good conscience...unless of course I was told I had to. The bottom line is you can build a model with PM and I can build it using the rules editor and there is no way in hades your model will be able to perform at the level mine will and without 10 times as much "generated" code. I'm sure PM will improve over time but I'm not going to hold my breath waiting on it and I'm certainly not going to give up the advantage I have over the Johnny-come-lately TM1 developers out there.

[qml]

I really expected this interesting thread to have picked up more speed by now (obviously, not in the original dierction). Let me add my two eurocents then.

First of all, I can't say I have worked with PM yet, so it would be unfair of me to critique it. However, I am sure that this tool hasn't been created as a response to popular demand from the TM1 developer community. In fact, Iboglix hasn't listened to us developers in a long time and this is unlikely to change. And so, @Mark, I would not expect to see a lot of jumping up and down on this forum about something that is, well, not a highly anticipated change, and as far as I can tell so far, not even a well developed one.

Is this the future of TM1? Probably. Will this sort of development interface increase the sales? Maybe. Will it mean that there will still be an ever-growing demand for experienced developers to come and clean the mess that the automatic code generation creates? Without a doubt.

Honestly, I'm not sure that PM is what this tool needed. TM1 has been a market leader in implementation times and maintenance costs anyway, even with the traditional, 19th-century-style programming interfaces, which, admittedly, needed a remake. Why fix what ain't broke if there are more important things waiting to be addressed? I agree, not everyone will see eye to eye on what the important things are. And I also agree that it's not the TM1 developers that are the money spenders, so it's no wonder we're not really being listened to.

As the last note, I'm with tomok on the assessment that TM1 is in its essence a planning/budgeting/forecasting platform and this is how it's been largely used since its birth. It has pretty much everything you can expect from a decent planning solution, including much better support for multiple writers in the last few releases (before the introduction of PI you really needed to know what you are doing in order to build a well-performing enterprise planning model).