Did you get these strings be converting an EXCEL spreadsheet into a SAS dataset? If so can you clean up the spreadsheet so it has only ONE type of value in each column? When converting a spreadsheet, where each CELL is independent, into a dataset, where each variable (aka column) must have only one type, then any column of cells that has mixed types must be converted to character to preserve the character strings.
The 2nd and 4th items are examples of how EXCEL will store TIME and DATE values and how SAS would then store them as character string when it has to store those numbers into a character variable.
Note that the FORMAT attached to the character is much less important than the storage LENGTH of the character variable. So let's just assume that for your four examples the LENGTH of the variables matched the displayed WIDTH that the format produces.
1) Arrival date: 01/02/2018 (length $11)
datevar = input(string,mmddyy10.);
If you really have any strings of length 11 then they might not work. You will want to look at those strings and see why they have a length of 11. Perhaps they have some extra spaces? In which case you might want to use COMPRESS() to remove those first so you don't ignore the last digit of the year.
If you want the first number to be the day of the month instead then use DDMMYY informat instead.
2) Arrival time: .4215277777777778 (length $21)
Just convert to a number and multiply by 24 hours. Remember the INPUT() function does not care if the width of the informat being used it longer than the string being read. 32 is the maximum width that the normal numeric informat will allow.
time_var = '24:00:00't * input(string,32.);
3)Radiology date/time: 01/04/2018 08:30 (length $19)
You can use the ANYDTDTM informat to read this, but you will want to set the system option DATESTYLE to the appropriate value so it knows whether to create the first of April or January fourth from that string.
options datestyle=MDY;
....
datetime_var = input(string,anydtdtm19.);
4) Read date: 43102 (length $10)
So that looks like it might be number of days since 1900, which how EXCEL normally stores dates.
date_var = input(string.32.) + '30DEC1899'd ;
Note you can also optionally ask EXCEL to count dates from 1904 instead which eliminates the goofy extra leap day issue. So if it looks like your dates are off by almost exactly four years try using this formula instead.
date_var = input(string.32.) + '31DEC1903'd ;
Remember to attach an appropriate FORMAT to your new numeric variables so they will print in a style that humans can understand. Avoid displaying dates in either MDY or DMY order has which ever one you pick will confuse half of your audience.
