Forum Discussion
First thing I did this morning was reviewed my proposed solution. This solution is feasible for narrow windows of time, which would be the assumption required to use it. However, there is a caveat on the century concern, which Kory brought up first, which is that if we’re reading a large window of time (for instance, say we’re trending inflation since WWII {I know, an extreme case, but makes the point}), then looking at “AUG-42” today using Date.parse() would result in “1942”, but if we process the exact same dataset this time next year, it will result as “2042” due to Date.parse() behavior around 19 vs. 20 years out. Very unreliable in that way.
So, use with caution.
Soapbox warning… I can’t figure out how we rounded Y2K and still have datasets that have 2-digit year representations. In my opinion, this should be fixed at the data source, where it is broken, and not rely upon fuzzy logic to try and figure it out by best guess. While solutioning this, I would have so many questions and caveats with which to annoy the requesting stakeholders.