Error using dtstr2dtvecmx. Failed on converting date string to date number.

Here is the script:

clear all; close all

clc

%%%%Directories

direc1 = '/Users/Degausseridae/Documents/Research/EggLoggerData/sas_00_createmetadata_Needs/';

direc2 = '/Users/Degausseridae/Documents/Research/EggLoggerData/sas_00_createmetadata_Produces/';

%%%%File to be read in

file2Load = 'BNSTdata.csv';

formatSpec1 = '%s %u %u %s %s %u %u %u %s %u %s %s %s %s %s %s %s %s %f %f %f %s %s %s %u %s';

formatIn = 'mm/dd/yyyy';

N = 26;

%%%%Import file

fileID = fopen([direc1,file2Load]);

C_data = textscan(fileID,formatSpec1,N,'Headerlines',1,'Delimiter',',');

fclose(fileID);

%%%%logger start date

sDate = datevec(C_data{1,13},formatIn);

Here is the error:

Error using dtstr2dtvecmx Failed on converting date string to date number.

Error in datevec (line 117) y = dtstr2dtvecmx(t,icu_dtformat);

Error in sas_00_createmetadata_GT (line 42)

sDate = datevec(C_data{1,13},formatIn);

Answer 1

[Walter Roberson]

At least one of the entries does not have the expected date format 'mm/dd/yyyy'. If you want to debug it then you could use:

C_dates = C_data{1,13};
ndates = length(C_dates);
sDate = -1 * ones(ndates, 6);    %using -1 to mark entries that did not convert

numwrong = 0;
maxwarnings = 10;      %maximum times to warn so we don't go overboard
for K = 1 : ndates
  try
    thisvec = datevec(C_dates{K}, formatIn);
    sDate(K,:) = thisvec;
  catch:
    numwrong = numwrong + 1;
    if numwrong <= maxwarnings
      fprintf(2, 'Warning: line #%d has date that is not mm/dd/yyyy, instead is |%s|\n', K, C_dates{K});
    end
  end
end

if numwrong > 0
   fprintf(2, '\nWarning: there were %d lines with invalid dates\n', numwrong );
end

This will, of course, be much less efficient than doing them all at once.

It would, by the way, be possible to only use the slow method if an error was found, by doing a try/catch on the mass conversion and only going one by one if something gets caught.

I probably wouldn't do it this way myself: I would probably use regexp() to match the valid date patterns. For example, if the strings might not have leading 0's on the month or day, then:

C_dates = C_data{1,13};
datepattern = '\d{1,2}/\d{1,2}/\d{4}';
nondates = cellfun( @isempty, regexp(C_dates, datepattern) );

and the problem entries are

C_dates(nondates)