Just wanted to list some TM1 tricks that we've learned from our time in the field:
- CREATING ALIASES – When creating aliases, use consistent attribute names. Name or Desc will usually suffice. (We frequently find that name is used on one dimension, desc is used on a couple others and description is used on a few more.) This reduces those “uh oh, what did I call that attribute” moments when creating a process or updating rules.
- IS A DYNAMIC SUBSET REALLY DYNAMIC? – Generally the extensive use of dynamic subsets is not recommended due to performance concerns. However, many subsets are dynamic only at the beginning of a cycle, not throughout the process. If the subsets are dynamic only at the beginning of a cycle consider creating them using a Turbo Integrator process.
- BUILDING PHASE SUBSETS – Views have a tendency to morph during the initial build phase and creating subsets can be time consuming if they are in flux, especially if one needs to add items back [since you must start over from the entire dimension list]. Set up an attribute on your row and/or column dimension that corresponds to the items that may be needed in a particular view; add a simple ‘Y’ to the attribute and create an attribute based subset that filters on the ‘Y’.
- UPDATING DIMENSIONS WITH A TURBO INTEGRATOR PROCESS -If there are multiple dimensions that use the same data source and are updated on a consistent basis, consider using only one TI Process. It is frequently possible to update an overall P&L dimension along with a Revenue account dimension and an Expense account dimension, especially when the Revenue or Expense Account dimensions require coding.
- LARGE CUBE OR SMALLER CUBES? Just because you can doesn’t mean you should. Yes, TM1 cubes can include as many as 256 dimensions but that doesn’t mean a model should consist of one cube with every dimension possible. Smaller cubes are both admin (rules/feeders are less cumbersome) and user friendly (easier to understand) as well as performance friendly.
Our list of TM1 tricks in Part 2 can be seen HERE.