An article from BATIMES popped up in my inbox this morning, the author raises the issue of varying date formats causing problems of misinterpretation. Different countries, different companies within the same country, different departments in the same company and different systems within the same department use differing formats. Even different reports from the same system use different formats (PlanView, I’m looking at you). Worst case I’m aware of is there’s a report I produce daily in a shared Excel spreadsheet where most people who use it want the DD/MM/YYYY format but one person insists on changing the format to M/D/YY whenever he looks at a tab.
The author of the article points out that we’re slipping back into the 2 digit year issue that cause problems with Y2K and that combined with different formats leads to her wonder if 11/05/18 is 11th May 2018 (DD/MM/YY), 5th November 2018 (MM/DD/YY) or 18th May 2011 (YY/MM/DD). Without a standard format, that is universally used, we can not be sure, especially going forward where the data may be separated from any accompanying documentation.
Looking at the datasets I have to deal with often the more fundemental problem is ensuring that users put only the data they are supposed to in a field. ‘N/A’ is not a valid date format in any country, nor is ‘Will provide next week’.
Stephen is learning data analysis 