Failed to read xml error when using xmlread

I am trying to read several xml files in a loop using xmlread. An error 'Failed to read xml file' occurs. On examining the xml file I noticed that in the first line that says <?xml version="1.0" encoding="ISO8859-1"?>, if I change ISO8859-1 to ISO-8859-1, xmlread works. Is there an automated way to corect this or any other way to read the files in bulk without having to manually correct the header in each file?

Answer 1

[dpb]

...
try
  DOMnode=xmlread(filename(i));     % try to read the file
catch ME                            % catch the failure; fixup
  fidi=fopen(filename(i),'r');      % open the file
  fido=fopen('tmp','r');            % open a scratch temp file
  while ~feof(fidi)
    l=fgetl(fidi);
    if ~empty(strfind(l,'ISO8859'))
      l=strrep(l,'ISO8859','ISO-8859'); % fixup the record
    end
    fprintf(fid0,l)                   % output to temp file...
  end
  fidi=fclose(fidi);
  fido=fclose(fido);
  copyfile('tmp',filename(i))          % and copy over the original
end 
DOMnode=xmlread(filename(i));       % and try again with corrected file...

[Sarah Immanuel]

Thanks a lot for your help.

Just a clarification: in this command fprintf(fid0,l) only the content of 'l' will be writtten to the tmp file? How do we get back all the other remaining content of the original file please?

[Walter Roberson]

It is within the loop, so eventually the entire content is written.

However, the

fprintf(fid0, l)

should be

fwrite(fid0, l)

Answer 2

[Walter Roberson]

filename = 'InputFileName.xml';
S = fileread(filename);
SS = regexprep(S, 'encoding="ISO8859-', 'encoding="ISO-8859-', 'once');
if strcmp(S, SS)
    remove = false;  %optimization, do not write new file if not needed
    tname = filename;
else
    tname = tempname();
    fid = fopen(tname, 'w');
    fwrite(fid, tname);
    fclose(fid);
    remove = true;
end

DOMnode = xmlread(tname);

if remove; delete(tname); end

This code is deliberate in narrowing down to encoding= and only doing the first instance, so as to avoid accidentally changing any ISO8859 that might happen to be part of the data.

[Sarah Immanuel]

Hi Walter, thanks a lot for your response.

I tried the above codes. To clarify: The replaced content is still within 'SS' and assuming strcmp(S,SS) is false, the 'tname = filename' is executed with filename still refering to the original (faulty) file with ISO8859 isnt it? How are the contents actually replaced within the faulty xml file please? Can you clarify this for me.

[Walter Roberson]

tname is set to filename when strcmp is true, not when it is false.

The comparison is true when the two strings S and SS are exactly the same, which would happen if regexprep did not make a change. Such as for a file that already has the right pattern, or which has a different encoding. In this situation the original file name is used directly for the later xmlread.

When the strcmp is false that means the original and regexprep versions are different, which means that the regexprep worked to make a new string. In that situation, a temporary file name is fetched, and the file is opened and the new content is written, and the temporary file is closed. It is this temporary file whose name is passed to xmlread. After the reading the temporary file is deleted

[Walter Roberson]

See also https://www.mathworks.com/matlabcentral/answers/101632-how-can-i-use-a-function-such-as-xmlread-to-parse-xml-data-from-a-string-instead-of-from-a-file-i#comment_972999 which shows a Java related method. To use it you would do the fileread(), regexprep(), and then java.io.StringBufferInputStream() the result, and xmlread() what you get from that.

Answer 3

[Sarah Immanuel]

Thanks a lot Walter, yes that makes sense. One last question, hope it is the last!. I am using Matlab2020a. The command tempfile() doesnt seem to work?

[Walter Roberson]

Sorry, should be tempname() instead of tempfile()

[Sarah Immanuel]

Hi Walter, thanks - the tempname() creates a tempfile but is not handled by the xmlread. It shows an error again. Can you help?

[Walter Roberson]

maybe

tname = [tempfile() '.xml'];