How to ask data transformation or automation questions
Ideally, you should include a sample initial data file as well as the desired result file. More advanced users should (when the data file is relatively complex) do this using
DATA LIST or
INPUT PROGRAM as this saves a lot of time to the person trying to answer the question. In some cases it took me as long to create the dummy data file as it took to solve the problem. The easier you make it for the potential solver, the greater your chances that he / she will devote time to help you out.
- Why: It is also useful to explain why you want to do this. The purpose of the why is that maybe there is better way to achieve your goal which does not require the described transformation.
- Number of cases & variables: The solution when one has 100 cases or 10,000,000 cases is not always the same. Similarly the number of existing variables will affect the solution. It is therefore advisable to provide that information.
- Frequency of use: An other element affecting the design of the solution is how frequently will the solution be applied. A one time use obviously requires less automation than a syntax which will be run every night through the Production facility.
In summary, you may simplify the example to make it easier to understand but to get a solution which really solves you problem, you should mention the elements listed above in order to described the full context.
Asking the right question remains the best strategy to get the right answer…
Students or teachers (who have access to huge libraries…) often ask questions of the type "Does anybody have a syntax to calculate statistics XYZ as described in Book ABC?". This is certainly a short way of asking a complex question but those of us who do not have readily access to a university library simply skip the question… If you have scanned pages of the formulas and could email them to persons interested, say so. Do not attach these documents to you postings.