Converting Mainframe Date Formats

Some of the data you work with in Excel may start as output from large systems in your office. Sometimes the date formats used by the large systems may be completely misunderstood by Excel. For instance, the output may provide dates in the format yyyydddtttt, where yyyy is the year, ddd is the ordinal day of the year (1 through 366), and tttt is the time based on a 24-hour clock. At first glance, you may not know how to convert such a date to something that Excel can use.

There are many ways that a solution could be approached. Perhaps the best formula, however, is the following:

=DATE(LEFT(A1,4),1,1)+MID(A1,5,3)-1+TIME(MID(A1,8,2),RIGHT(A1,2),0)

This formula first figures out the date serial number for January 1 of the specified year, then adds the correct number of days to that date. The formula then calculates the right time based on what is provided.

When a formula like this is invoked, the result is a date serial number. This means that the cell still needs to be formatted to display a date format.

This approach will work just fine, provided that the information that you start with makes sense. For instance, you will always get the expected result if ddd really is within the range of 1 through 366, or if tttt is a properly formatted 24-hour representation of time. If you anticipate original data that could be out of bounds, then the best solution is to create a custom function (using a macro) that will tear apart the original data and check the values provided for each portion. If the data is out of bounds, the function could return an error value that would be easily detected within the worksheet.