Convert different datetimes into one format

Hi everyone,

I'm having 2 different date types (1300 rows) in one column (see below). How can I convert them into one date form. I tried different suggestions, it worked only for one type the other one showed NaT. I would prefer (xx/yy/zzzz). I appreciate your help.

'03-Sep-2009'
'17/3/2009'
'05-Feb-2009'
'05-Apr-2009'
'05-Sep-2009'
'18/5/2009'
'25/5/2009'
'28/5/2009'
'28/5/2009'
'28/5/2009'
'06-Feb-2009'
'06-Mar-2009'
'06-Apr-2009'
'06-Apr-2009'
'06-Apr-2009'
'06-Apr-2009'
'06-Apr-2009'
'06-Apr-2009'
'06-Apr-2009'

Answer 1

[dpb]

Bestest would be to fix the input source to use the same format but you can try some subterfuges to see if you can sneak by without having to parse every one individually to determine which format it is on its own. datetime will do that on its own (at a high overhead so it may get slow), but you can try--with the list above as cellstr array:

>> cellfun(@(s)datetime(s,'format','dd\MM\yyyy'),dstr)
ans = 
  19×1 datetime array
   03\09\2009
   17\03\2009
   05\02\2009
   05\04\2009
   05\09\2009
   18\05\2009
   25\05\2009
   28\05\2009
   28\05\2009
   28\05\2009
   06\02\2009
   06\03\2009
   06\04\2009
   06\04\2009
   06\04\2009
   06\04\2009
   06\04\2009
   06\04\2009
   06\04\2009
>> 

If you get one of the ambiguous forms like 01\01\2009 you'll get a warning and will have to determine whether the default logic is correct or not...unfortunately, I don't think datetime tells you which particular entry(ies) was(were) the bad one(s)

Alternatively, do the first thing as passing the array, then locate the isNAT locations and reprocess them with the proper input format and place them where they belong. That's a loop for as many distinct input forms there are.

All in all, fixing the source might still be the easier and surely would be the better way.

[Ha Sen]

Great! Thank you.