Strange error using readtable

Greetings,

Having trouble with readtable with text files. The text files in question have a combination of text and number, thus the reason for using readtable. The script runs through the first two text files with no issues then crashes on the first line of the next text file.

while row <= height(workfile) %Height command identifies the number of rows in the table to stop the loop.
        
        t=workfile(row, col);  % t is a temp variable to store the date before processed by datenum
        t=table2array(t); % convert to an array
        t=datenum(t);  % process to a datenum value
        
        col=3; % set column to 3 to pull precip data
        
        % now pull the precip data
        p=workfile(row, col);
        p=table2array(p);
        
        Vdata=[t, p]; %temp variable to allow easier use of the fid process
        
        fid=fopen (verfile , 'a');
        fprintf(fid, '%6.0f, %7.4f \n ' , Vdata);
        fclose(fid);

The datenum line executes fine but it appears the issue comes up with the fprintf command as I get this error

Error using fprintf
Function is not defined for 'cell' inputs.

Error in sitedata (line 95)
        fprintf(fid, '%6.0f, %7.4f \n ' , Vdata);

LIke I said, this was not an issue with the previous text files that were sucessfully run before so I am stumped. Below is a sample of the text files in question:

10/31/2018	0	0
11/1/2018	0	0
11/2/2018	0	0
11/3/2018	0	0
11/4/2018	 T	0.0001
11/5/2018	0	0
11/6/2018	0	0
11/7/2018	0	0
11/8/2018	 T	0.0001
11/9/2018	0.19	0.19
11/10/2018	0	0

This is what the file before the one the crashed look like (top of the file).

Here is the one the script crashed on (same area of the file):

10/31/2018	0	0
11/1/2018	0.1	0.1
11/2/2018	0.45	0.45
11/3/2018	0	0
11/4/2018	0.5	0.5
11/5/2018	0.14	0.14
11/6/2018	0.11	0.11
11/7/2018	0	0
11/8/2018	0	0
11/9/2018	0.12	0.12
11/10/2018	0	0

Other than the data, they don't look any different. Any ideas??

[Walter Roberson]

Could you confirm that the files that work properly have 'T' entries in some rows, and that the files that do not work do not have 'T' entries? I would have expected the other way around.

Or perhaps there are text entries in the part you did not show us in column 3 of the file that does not work but no text entries in column 3 of the ones that work?

If there are text entries similar to the 'T' we can see in the example then what do you want done with them?

[Eric Metzger]

Yes, the "T" entries work fine in the other files. Here is an example of a previous file that ran with a "T" value.

10/31/2018	0	0
11/1/2018	0	0
11/2/2018	0	0
11/3/2018	0	0
11/4/2018	 T	0.0001
11/5/2018	0	0
11/6/2018	0	0
11/7/2018	0	0
11/8/2018	 T	0.0001
11/9/2018	0.19	0.19
11/10/2018	0	0
11/11/2018	0	0
11/12/2018	0.2	0.2
11/13/2018	 T	0.0001

I have looked over he file that consistently fails with this error. I can't find anything different about it. I have attached the full file as a reference.

In this script, the second colum is ignored as it was processed separatly by excel. Only columns 1 and 3 are used.

Here is the more of the code and the complete "while" loop for this part. Hopefully this explains it better.

D=dir('./*pre.txt');  % Identify the files to be processed

for i=1: length(D)
    file=D(i,:) .name; % pull the file to be processed
    workfile=readtable(file);  % puts the file in a format that can be processed


% pull the date data out and reformat into a datenum

% variables for rows, columns and the lenght of the file

row=1;
col=1;
vfile=file(1:5);  % pull the file name out for the complete verification file name
verfile=[vfile, 'ver.txt'];  % make the output file name
station=file(1:3);  % Identifies the verification site being worked with to set up the 72hr verification file
pfile=['Processing file:  ' file];
disp(pfile)
precip72hrfile=[station, '72hrver.txt'];  %72hr precipitation file

while row <= height(workfile) %Height command identifies the number of rows in the table to stop the loop.
        
        t=workfile(row, col);  % t is a temp variable to store the date before processed by datenum
        t=table2array(t); % convert to an array
        t=datenum(t);  % process to a datenum value
        
        col=3; % set column to 3 to pull precip data
        
        % now pull the precip data
        p=workfile(row, col);
        p=table2array(p);
        
        Vdata=[t, p]; %temp variable to allow easier use of the fid process
        
        fid=fopen (verfile , 'a');
        fprintf(fid, '%6.0f, %7.4f \n ' , Vdata);
        fclose(fid);
        
        row=row+1; %move to the next row
        col=1; %reset the column to for the next time through the loop.
        
end

[Eric Metzger]

Hmmm...i see. I guess this is one that got by me. One of hte artifact errors the excel script was supose to identify and strip out. Will have to check.

To be sure, if there are letters and number like one Walter identified, will this make the "table2array" command fail?

If so,this could be why the table stayed a string and not conver to an array.

[per isakson]

In brtyrpre.txt the pair

5/30/2019	 S	 S
5/31/2019	 0.22A	 0.22A

appears twice. In your question

11/4/2018	 T	0.0001

appears twice. Will other letters pop up?

The construct, which reads a value from the table

x = workfile( row, col );  
x = table2array( x ); 

can be replaced by

x = workfile{ row, col }; 

"To be sure, if there are letters and numbers ..." See the two <3x3 table> in my answer and notice that one contains doubles the other characters. And read my text

With letters among the numerical data the entire column is imported as character, without letters it's imported as double. When "extracting" a character column from a table you get a cell array, which causes the error.

table2array doesn't fail!

Answer 1

[per isakson]

I've downloaded brtyrpre.txt from your comments and copy&pasted brtyr_error_pre.txt from your question ("Here is the one the script crashed"). Next I run the script of your comment. brtyr_error_pre.txt was processed without problems and the script throw an the error for the first row of brtyrpre.txt.

>> Untitled
Processing file:  brtyr_error_pre.txt
Processing file:  brtyrpre.txt
Error using fprintf
Function is not defined for 'cell' inputs.
Error in Untitled (line 34)
        fprintf( fid, '%6.0f, %7.4f \n ' , Vdata ); 
K>> 

The the text of the error differs somewhat, but the message is the same. I run R2018b.

The reason for the error is that in one of the files there are letters among the numerical data (as Walter assumed). With letters among the numerical data the entire column is imported as character, without letters it's imported as double. When "extracting" a character column from a table you get a cell array, which causes the error.

"The script runs through the first two text files with no issues then crashes on the first line of the next text file." Run one file at a time to become 101% sure which file causes the error.

K>> workfile(1:3,:)
ans =
  3×3 table
       Var1       Var2    Var3
    __________    ____    ____
    10/31/2018       0       0
    11/01/2018     0.1     0.1
    11/02/2018    0.45    0.45
K>> workfile(1:3,:)
ans =
  3×3 table
       Var1        Var2      Var3 
    __________    ______    ______
    10/31/2018    '0'       '0'   
    11/01/2018    '0.1'     '0.1' 
    11/02/2018    '0.45'    '0.45'    

And see

• Debug a MATLAB Program and
• Examine Values While Debugging

[Walter Roberson]

In particular you have 0.25A and 0.22A

Note: there is a Suffixes option that can be set against the results of detectImportOptions to remove known suffixes, in the case where 0.25A is to be treated as if 0.25 were present instead.

[Eric Metzger]

Thanks to both of you (Walter and Per). I am not very proficent with MatLab but it is the code that is much easier for me to manipulate and get the results I want. I still have problems with little obviouse stuff like this.

Thanks again!