Stata tip 98: Counting substrings within strings

摘要

Consider the following problem, based on a real one reported on Statalist. A user has string date-times in differing forms; let us imagine they are in a variable, sdate. The desire is for everything to be DMY hms-that is, day, month, year, space, hours, minutes, and seconds. Examples of data are "25/12/2010 11:22" "25/12/2010 11:22:33" "25/12/2010 11:22:33:444" Compared with the standard, the first example lacks seconds; the second example is fine; and the third example includes milliseconds, but expressed in a nonstandard way. Stata expects seconds and milliseconds in the decimal form "33.444".