how to speedup the function datetime('('1/1/2014 1:00:00 AM','Conve​rtFrom','e​xcel')?

I have a number of date-time values stored in an excel file. The format is '1/1/2014 1:00:00 AM'. I need to convert it into matlab date-time values, so I use the function datetime(). But the trouble is I need to to convert a number of such values, say 10,000, the runtime is rather long. So could I speed up this procedure？ I grateful for any help

[Walter Roberson]

When you read them from the excel files, are they appearing as strings or are they appearing as numeric values ? Excel typically stores dates as floating point values so that is more likely.

Answer 1

[Are Mjaavatten]

 [num,txt,raw] = xlsread('file.xlsx')

returns date-time values as NaNs in the num output and as strings in the txt and raw outputs. I often find it useful to add a new date-time column in the Excel sheet, copying the original, but formatted as a number. The num output will contain this column as numeric values. To convert the date values in column 1 of num from the Excel date format to Matlab:

 d = num(:,1) + datenum(1899,12,30);

[Walter Roberson]

I have definitely seen excel date-time values come out as numeric. It might depend on whether you are using Windows with Excel installed or are using 'basic' mode (everything else)

[Are Mjaavatten]

It seems that Walter's comment is correct unless you are using Windows. And even on Windows, there is a simpler workaround than the one I proposed:

 num = xlsread(file,'','','basic');

will return datetime values as numbers. Convert to Matlab's date standard as indicated in my answer above.