About DavidSmith

DavidSmith

Agreed - this is time consuming, but in the meantime, this is the process I use, using an old Excel trick in learnt!! Export the line items to Excel It is unlikely that Column J is used so it is a good one to use, otherwise use another appropriate column that follows the same pattern of nothing for header Filter on blanks Click on J2 and enter the formula =a2 and copy this down to the rest of the header lines Remove the filter Now highlight from J2 down to the last row in the blueprint From the “Goto” menu option (usually <F5>), choose Special and then Constants This should select all of the “detail” line items Without clicking anywhere else, type =J2 and then CTRL+Enter, or CMD+Enter This will enter the return the module name against each line item Voila!

DavidSmith

@david.savarin I have run the analysis using the above model and set formula scope for Current Version only For the Fake versions model I used a single boolean to check and had a formula like this: IF Fake Version Details.'Not Current?' THEN 0 ELSE CUMULATE(Data) The results were almost exactly the same as above The volume of the calculations for Fake Versions were 2x larger, but the duration of the calculations were 6x shorter The checking for the "not current" prevents the calculation going further (a benefit of the early exit condition) I also checked the effect of the early exit (checking for the most common condition of not current vs current) As expected, checking for the IF "Not Current" THEN 0 ELSE ... was 1.5x faster than IF "Current" THEN Cumulate() ELSE 0 One other element to add to the overall discussion is size and switchover. Using Switchover will reduce the model size, as the "actual" periods don't take up space. So if size is a primary consideration, bear that in mind. As ever, if is a balance between performance, size, sustainability, usability etc. I hope that helps David

DavidSmith · ‎12-20-2019

@ArunManickam Thank you

DavidSmith · ‎12-18-2019

Thanks @adam_bimson_5 And yes, Versions as a format is on the list as @stephen.brook said David

DavidSmith · ‎12-17-2019

Calculating the difference between two dates is a common requirement and one that you will likely need to do at some point. Below you will find the best practice for calculating the difference between two days, and the best practice for calculating the difference between two months. While there are multiple ways to get to the answer you're looking for, often the simplest solution is the best. When calculating the difference between two dates (days or months), use this approach first: For date formatted line items, use: line item 1 – line item 2. For time formatted line items, use: START(line item 1) – START(line item 2). Dates 1. Simple dates difference. 2. Dates addition. 3. Dates subtraction. Time Periods 1. Time period difference. 2. Months addition. 3. Months subtraction. 4. Years addition. 5. Years subtraction. Keep 'Superset' in Mind When Selecting Your Time Period If you choose a time period that is outside of the superset, it will display as blank. The superset is the earliest date from the model calendar and any time range through to the later of the last date of the calendar of any time range. Example model: Time: Superset = Jan 17 - Dec 22: If you go outside of that range: You're now ready to calculate date differences in your daily practice! If you have additional questions or feedback, please add them to the comments below.

DavidSmith · ‎12-13-2019

Thank you all for the kind words. It is very humbling I'm really excited about the opportunity to make Anaplan even better and having the support internally and externally makes it an easier transition It is going to be difficult moving away from modelling; something I have been doing for the last 30 years, but my intellectual challenge now moves from creating workarounds, to help to build a product so that you don't need so many workarounds I am going to miss the community spirit, but I am still here (in the background), I will attend CPX, and other events as appropriate and I will also be wanting to engage community members and Master Anaplanners to help with the development of the platform Cheers David

DavidSmith · ‎12-13-2019

All I think this is a bug, because it should work in line with how subsets work I'll raise it and get it on the backlog to be investigated David

DavidSmith · ‎12-12-2019

Unless anything has changed, it's all about the Unicode base plane https://en.wikipedia.org/wiki/Plane_%28Unicode%29 I have no idea what all that means, but when Michael told me not to do it, I listened David

DavidSmith · ‎12-12-2019

One of my initial investigations/priorities is around sorting and the list order. There is obviously a "natural" order embedded in there; that is how CUMULATE on a list works, but I want to explore the implications of resetting that, or using a defined order instead David

DavidSmith · ‎12-12-2019

@dameyer01 The only thing you can try is to check the name of the source file (from the import sources tab). If you can find an existing file that "might" be what was used, load that You may then be able to see some of the mappings I hope that helps David

DavidSmith · ‎12-11-2019

As you may know by now, I have taken on a new role in the Product organisation of Anaplan. I've taken responsibility for the evolution of the Core Platform. This will be quite a departure for me, having worked for 20 years on the "other side of the fence" but I'm looking forward to brining all of my experience to date of what Customers really need to the development of our platform. So that can focus fully on the new task, I am going to have to relinquish my community duties, so from next week, I will be effectively "retiring" from community. I will publish V2 of the Planual next year and I will be reviewing the Idea Exchange and all of the suggestions to prioritise the development path for the next year. I will also be engaging community members to assist in that prioritisation exercise in due course. So I'm not going away entirely!! What I would say though is I'm happy that I'm leaving the community ecosystem in a great place. The progression in the last year has been amazing and I'd like to call out a few individuals who have championed best practices and supported the Anaplan ecosystem. Keep it up! So thank you to: @rob_marshall @nathan_rudman @LipChean_Soh @ArunManickam @usman.zia @Misbah @Jared Dolich @kavinkumar @vignesh Happy modelling my community friends David

DavidSmith · ‎12-11-2019

@Misbah Especially for you!!!

DavidSmith · ‎12-11-2019

@Jared Dolich There is always the "why" question Why do we need to that complexity in the mappings?. when you resolve through the patterns you often can rationalise and simplify David

DavidSmith · ‎12-11-2019

@MaoNinomiya Apologies for the delay, but I think I have a solution. It is complex and a little messy, so I have included it in the attached. It seems to have satisfied your example data, but obviously there are a lot of patterns missing. However, hopefully it shows you some principles of how to break down the mapping formula into manageable and testable pieces!! I hope it helps David

DavidSmith · ‎12-09-2019

@jim_ma exactly. Glad it solves it for you David

DavidSmith · ‎12-09-2019

@Sameer_Bajaj Exactly - spot on,, where State is a line item formatted as State in a module dimensioned by Medicine David

DavidSmith · ‎12-09-2019

@jim_ma It is a bit of a misconception that you users to enter data in a module they need write access to the list. They don't. Access to the module will suffice One only needs to grant write access to a role if that role needs to add, amend or delete members from the list So as @Misbah correctly stated, in most cases (and especially if being sourced from a data hub), it is best practice not to give roles access to the list. Not only does this reduce the risk of "accidental" corruption, but it is also more efficient as the system doesn't need to pre-allocate memory for the next block of "inserts" on the list The exception is, of course, if adding items is a part of the user/role process (e.g, promotional, scenario or project planning) David

DavidSmith · ‎12-08-2019

@Jared Dolich Yes However, you cannot do it on one go First you need to create the copy and then move to the desired target workspace and import the new copy into there The golden rule of compatible models is that the last revision tag in the target must exist in the source. Assuming there are no outstanding structural changes in the model to be copied, when it is copied, it creates a revision tag in itself and the new copy, thus the compatibility is set up. Importing the new copy to the new workspace doesn't set another revision tag, so the compatibility is maintained I hope this helps explain it David

DavidSmith · ‎12-08-2019

@Sameer_Bajaj From your initial diagram it looks like all Medicines belong to each state, so a module dimensioned by State and Medicine will suffice; you don't need a LOOKUP at all. A module dimensioned by State to hold the %s and then in the target module, you can just reference the previous line item If you have a relationship between State and Medicine, then you do need to model the relationship between State and Medicine One of the big advantages of Anaplan is we do not always need parent child relationships to aggregate or model, as previously described. If there is only one State for any one Medicine, then you can just have a module dimensioned by Medicine and have a line item formatted as State to hold the mapping. then use a LOOKUP to the State %s module If these concepts are difficult for you to understand, then I would suggest you take the Level 1 training, if you haven't already and also look at the attached. Both explain how to use SUM and LOOKUP I hope this helps David

DavidSmith · ‎12-06-2019

For me the other big areas that make Anaplan different are: Anaplan ensures consistency of data for whomever you are, whatever you are looking at, at whatever level Anaplan doesn't need to use a single set of structures throughout the model. We can use "appropriate" structures to ensure efficiency of calculation (e.g DISCO) Anaplan can aggregate without hierarchies Finally, we do have a super optimised engine and use of computer processing that allows the rapid calculation performance at scale And I'm proud to work for such a company David

DavidSmith · ‎12-06-2019

@T_Caban It's not really a performance issue (although it probably would be!), it is just not the way Anaplan works. Everything has to be linked back to a specific object. Text is a free form data point, so can't be used to tie into structures David

DavidSmith · ‎12-06-2019

The data is stored/calculated at the intersection of the different dimensions (Lists, Time, Versions) Text values are not part of the list definition (Name and Code) are, so that is why we use formatted line items, and use FINDITEM to return the list member for use in modelling It is not the same structures as Excel, which uses generic rows and columns. Anaplan is much more structured than that It would be easier not to have to use FINDITEM to match/transform, but that is a whole different topic!! I hope that explains it David

DavidSmith · ‎12-05-2019

@tatsumi We've had a number of similar threads on this (which I'll try and find), but the easiest way is to add a column to your source data file that is formatted as "TRUE" In your model (ideally Data Hub), in the System module for the data in question, you have two line items (e.g. Reset? and Imported?). Reset? is a set to No Data and Imported? is a Boolean As part of the import process: first run an internal import from Reset? to Imported? Then when you bring in the data, populate the Imported? line item with the extra column from the source file That shows exactly which records came in as part of the import. You can then filter on these if necessary, or add additional functionality for downstream exports if required I would add that we are going to add an example of this as part of our Level 3 / Solution Architect training I hope that helps David

DavidSmith · ‎12-05-2019

@Tom_Brumfield Unfortunately, that setting is not able to be changed. The "organisation' list is an autogenerated list when a model is created and the list cannot be deleted. The only way to now align with the Level 1 training is to rename the list and use it in the hierarchy, but if you have already created the hierarchies, personally, I wouldn't bother. The check box is a hang over from way back and as far as I know it is not used for anything, and it definitely won't affect your ability to pass Level 1 Good luck with the training David

DavidSmith · ‎11-27-2019

If you have any questions on the planual, please post them in the comments on the Planual page

DavidSmith · ‎11-26-2019

@CallumW We are looking at this, it's in progress, so hopefully we'll have something out soon David

DavidSmith · ‎11-25-2019

Sometimes you have to be pragmatic - Similarly if you are using a "delete from list" module - I often have one of those with many line items as Subsidiary views Like we say about the Planual rules, there are, and always will be exceptions DAvid

DavidSmith · ‎11-25-2019

@dameyer01 Not as such 1. Create a module dimensioned by the subset and create a single boolean line item and set it to TRUE 2. Then in your main list module, create a boolean line item (corresponding to the subset) and reference the line item from above. this module should be your System module for the list, not dimensioned by anything else 3. That can then be used to know if the list item is in the subset If you have a lot of subsets, then it might be practical to create a single module and use subsidiary views (this is one of the exceptions to the rule) for each subset I hope that helps David

DavidSmith · ‎11-25-2019

The challenge you have here is the line item is dimensioned by Account L1 and in a composite hierarchy Account L0 is a calculation So we need to build a couple of extra line items 1. You need to know if the item is a L1 or L0, so set up a Boolean line item called Child? and set this to TRUE 2. Create another line item that brings in the Parent of the L1 - in your model it will be something like PARENT((ITEM('Account L1')). It is always best to have this in a separate line item (You will likely use it again). Remember "calculate once, reference many times" 3. Create a text line item for the name of A0, but set the summary option to 'Last non-blank' 4. This is the existing line item we just solved for 5. Now you bring it all together with a conditional and set this summary option to 'Formula' Viola! David

DavidSmith · ‎11-25-2019

In a word - Messy! To avoid a series of IFs, I would, instead, create a series of boolean modules dimensioned by Code to be mapped and the different rule logic for the different codes/number patterns. Some do not apply, as per your example, Code C, D ,E and F are not applicable for Mapped Codes A-F, so you would set those to TRUE. You will probably need line items for each pattern, but splitting it up this way is the most efficient, but you will need to update the formulas/add more Can you post some example data to help understand the formats of Code A - Code F patterns? It's always easier to work with data than the theory David David