Time series data formatting issues

I have 15 minute interval time series data that is in the format:

MM/dd/yyyy hh:mm:ss a

However at 12:00:00 on each day the time stamp is recorded as

MM/dd/yyyy

with no hh:mm:ss a.

Therefore when I convert it from cell data to datetime, the values for the new day is recorded as a NaN.

How can I get each of the midnight "NaN"'s to record as MM/dd/yyyy hh:mm:ss a?

Thanks

Answer 1

[Peter Perkins]

Clay, the simplest thing to do is just fill in the NaTs after the fact:

>> c
c =
    '12/01/2016 10:00:00 PM'
    '12/01/2016 11:00:00 PM'
    '12/02/2016'
    '12/02/2016 01:00:00 AM'
    '12/02/2016 02:00:00 AM'
>> d = datetime(c,'Format','MM/dd/yyy hh:mm:ss a')
d = 
   12/01/2016 10:00:00 PM
   12/01/2016 11:00:00 PM
   NaT                   
   12/02/2016 01:00:00 AM
   12/02/2016 02:00:00 AM
>> d(isnat(d)) = datetime(c(isnat(d)),'Format','MM/dd/yyy')
d = 
   12/01/2016 10:00:00 PM
   12/01/2016 11:00:00 PM
   12/02/2016 12:00:00 AM
   12/02/2016 01:00:00 AM
   12/02/2016 02:00:00 AM

You could also find the "short" strings and append '00:00:00 AM' to them before calling datetime.

Hope this helps.