Besides being slightly unconventional, are there any downsides with having a bunch of empty globals in a table?
Reason: we have a number of solutions and several of them supply data on a semi-regular basis to another agency. They import this data into their database. Each of our solutions does different things, so naturally the architecture varies between them. The external agency has a new tool for importing our data, plus data from other entities like us. Thus, it's helpful to them if all the data is supplied in a consistent csv/ Excel sheet, with the same number of columns and specific fields being in predictable - and consistent - places (across the external suppliers). They have provided us with their 'wish list template', to see if we can accommodate their preferences.
It seems to me that the easiest way to achieve this is for us to add 'placeholder' fields - 20-40, depending upon the solution - so that we can deliver 'template-style' datasets as required.
Those placeholder fields would be formatted as non-enterable globals, 'number' for good measure.
While it doesn't conform to the ideal of narrow tables, it seems to me that it will be a low overhead way to collaborate on a practical level. Are there any 'gotchas' that I'm overlooking?