delete lines in txt based off phrase

I would like to delete sections of a txt file based off a phrase. so i want to find NaN as seen bewlow then delete anythign around that phrase to the { then below to the }, . so from origonal file to desired file. my code is attached. i can manage to delete NaN but nothing else.

origonal

{
    Name = PTC_FAILURE_CRITERION_TYPE
    Type = String
    Default = 'NONE'
    Access = Locked
    },
{
    Name = PTC_TENSILE_YIELD_STRESS
    Type = Real
    Default =     NaN psi
    Access = Full
    },
{
    Name = PTC_TENSILE_ULTIMATE_STRESS
    Type = Real
    Default =     NaN psi
    Access = Full
    },
{
    Name = PTC_MASS_DENSITY
    Type = Real
    Default = 1.000000 lbm/in^3
    Access = Full
    },
{
    Name = MATL_TYPE
    Type = String
    Default = 'UNASSIGNED MATERIAL'
    Access = Full
    },

desired

{
    Name = PTC_FAILURE_CRITERION_TYPE
    Type = String
    Default = 'NONE'
    Access = Locked
    },
{
    Name = PTC_MASS_DENSITY
    Type = Real
    Default = 1.000000 lbm/in^3
    Access = Full
    },
{
    Name = MATL_TYPE
    Type = String
    Default = 'UNASSIGNED MATERIAL'
    Access = Full
    },

%working
       fid = fopen('trial.txt', 'rt')  % Open source file.
       g= fscanf(fid,'%c');
      % gc=convertCharsToStrings (g)
       missing =strfind(g,'NaN')      
             numg= strlength(g);
 
       fid2 = fopen('trial.mtl', 'w+');
for i = 1:1
    if isempty(missing)
         full=g;
        fprintf(fid2,'%s', g);
    elseif missing>=true  %
         %delet NaN
       
         expression = '.NaN';
        replace = '';
       newStr = regexprep(g,expression,replace);
     
       
       %delet area 
       %these are my tries (deosnt work)
       %expression2 = 'Name = PTC_POISSON_RATIO' %([\n\r]+\  Type = Real[\n\r]+) Default =
      % expression3= 'Access ='
       %expression2 = ' </?{}.*?>'
      % replace2 = 'blank';
     % newStr1 = regexprep(newStr,{expression2, expression3}, {replace2, replace2})
             %expression = 'Default(\w+)=';
           %newStr1=  strrep(newStr,expression2,replace2);
       % replace = '';
      % newStr1 = regexprep(newStr,expression,replace)
                   
         
    end
    finalword=newStr1;
end
fprintf(fid2,'%s', finalword) ;
fclose(fid2);
        fclose(fid);
   
       

[dpb]

I'd read in the whole file into a cellstr or string array, then find the locations containing the section header record, "Name" and the wanted phrase "NaN"

Then locate the indices in the idxNAN array that are between the idxHeader locations -- those give you the bounds for the cell locations/records to delete.

NB: that this can be done globally for the initial indexing but will have to process from top down to avoid changing locations when deleting sections. That process is a loop over the number of elements in the idxNAN array.

[Kristin Bogar]

ok thank you!

[dpb]

See Answer I posted probably while you were posting this... :)

Answer 1

[dpb]

txt=textread('bogar.txt','%s','whitespace','','delimiter',newline); % read file a cellstr array
idxHdr=find(contains(txt,'Name'))-1;                                % find the section headers
idxTrlr=find(contains(txt,'},'));                                   % and the closing trailers
idxNaN=find(contains(txt,'NaN'));                                   % then offending NaN locations
ixDel=flip(interp1(idxHdr,1:numel(idxHdr),idxNaN,'previous'));      % find sections containing NaN
for i=ixDel(:).'                                                    % iterate over sections
  txt(idxHdr(i):idxTrlr(i))=[];                                     % remove between header/trailer
end

Applied to your example file, the above produces:

>> txt
txt =
  18×1 cell array
    {'{'                                    }
    {'    Name = PTC_FAILURE_CRITERION_TYPE'}
    {'    Type = String'                    }
    {'    Default = 'NONE''                 }
    {'    Access = Locked'                  }
    {'    },'                               }
    {'{'                                    }
    {'    Name = PTC_MASS_DENSITY'          }
    {'    Type = Real'                      }
    {'    Default = 1.000000 lbm/in^3'      }
    {'    Access = Full'                    }
    {'    },'                               }
    {'{'                                    }
    {'    Name = MATL_TYPE'                 }
    {'    Type = String'                    }
    {'    Default = 'UNASSIGNED MATERIAL''  }
    {'    Access = Full'                    }
    {'    },'                               }
>> 

rewrite to a file... writecell would be good for that.

[dpb]

idxHdr=find(contains(txt,'Name'))-1;                                % find the section headers

Actually, I notice that the file doesn't have embedded "{}" blocks so one could use

idxHdr=find(contains(txt,'{'));                                % find the section headers

instead and eliminate the offset from the 'Name' record. It's a nit, but...