Mercurial > hg > camir-aes2014

annotate toolboxes/FullBNT-1.0.7/KPMtools/loadcell.m @ 0:e9a9cd732c1e tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
rev   line source
wolffd@0 1 function [lc,dflag,dattype]=loadcell(fname,delim,exclusions,options);
wolffd@0 2 %function [lc,dflag,numdata]=loadcell(fname,delim,exclusions);
wolffd@0 3 %
wolffd@0 4 % loadcell loads a cell array with character delimited
wolffd@0 5 % data, which can have variable length lines and content.
wolffd@0 6 % Numeric values are converted from string to double
wolffd@0 7 % unless options is a string containing 'string'.
wolffd@0 8 %
wolffd@0 9 % loadcell is for use with small datasets. It is not optimised
wolffd@0 10 % for large datasets.
wolffd@0 11 %
wolffd@0 12 % fname is the filename to be loaded
wolffd@0 13 %
wolffd@0 14 % delim is/are the relevant delimiter(s). If char(10) is included
wolffd@0 15 % newlines are simply treated as delimiters and a 1-d array is created.
wolffd@0 16 %
wolffd@0 17 % exclusions are the set of characters to be treated as paired
wolffd@0 18 % braces: line ends or delimiters within braces are ignored.
wolffd@0 19 % braces are single characters and any brace can pair with
wolffd@0 20 % any other brace: no type pair checking is done.
wolffd@0 21 %
wolffd@0 22 % options can be omitted or can contain 'string' if no numeric
wolffd@0 23 % conversion is required, 'single' if multiple adjacent seperators
wolffd@0 24 % should not be treated as one, 'free' if all linefeeds should be stripped
wolffd@0 25 % first and 'empty2num' if empty fields are to be treated as numeric
wolffd@0 26 % zeros rather than an empty character set. Combine options using
wolffd@0 27 % concatenation.
wolffd@0 28 %
wolffd@0 29 % lc is a cell array containing the loaded data.
wolffd@0 30 %
wolffd@0 31 % dflag is a set of flags denoting the (i,j) values where data was entered
wolffd@0 32 % dflag(i,j)=1 implies lc(i,j) was loaded from the data, and not just set
wolffd@0 33 % to empty, say, by default.
wolffd@0 34 %
wolffd@0 35 % numdata is an array numdata(i,j)=NaN implies
wolffd@0 36 % lc(i,j) is a string, otherwise it stores the number at i,j.
wolffd@0 37 % This will occur regardless of whether the 'string' option is set.
wolffd@0 38 %
wolffd@0 39 % lc will return -1 if the file is not found or could not be
wolffd@0 40 % opened.
wolffd@0 41 %
wolffd@0 42 % Hint: numdata+(1/dflag-1) provides a concise descriptor for the numeric data
wolffd@0 43 % Inf=not loaded
wolffd@0 44 % NaN=was string or empty set.
wolffd@0 45 % otherwise numeric
wolffd@0 46 %
wolffd@0 47 % EXAMPLE
wolffd@0 48 %
wolffd@0 49 %[a,b]=loadcell('resultsfile',[',' char(9)],'"','single-string');
wolffd@0 50 % will load file 'resultsfile' into variable a, treating any of tab or
wolffd@0 51 % comma as delimiters. Delimiters or carriage returns lying
wolffd@0 52 % between two double inverted commas will be ignored. Two adjacent delimiters
wolffd@0 53 % will count twice, and all data will be kept as a string.
wolffd@0 54 %
wolffd@0 55 % Note: in space-separated data 'single' would generally be omitted,
wolffd@0 56 % wheras in comma-seperated data it would be included.
wolffd@0 57 %
wolffd@0 58 % Note the exclusion characters will remain in the final data, and any data
wolffd@0 59 % contained within or containing exclusion characters will not be
wolffd@0 60 % converted to numerics.
wolffd@0 61 %
wolffd@0 62 % (c) Amos Storkey 2002
wolffd@0 63 % v b160702
wolffd@0 64
wolffd@0 65 % MATLAB is incapable of loading variable length lines or variable type values
wolffd@0 66 % with a whole file command under the standard library sets. This mfile
wolffd@0 67 % fills that gap.
wolffd@0 68 if (nargin<4)
wolffd@0 69 options=' ';
wolffd@0 70 end;
wolffd@0 71 dflag = [];
wolffd@0 72 %Open file
wolffd@0 73 fid=fopen(fname,'rt');
wolffd@0 74 %Cannot open: return -1
wolffd@0 75 if (fid<0)
wolffd@0 76 lc=-1;
wolffd@0 77 else
wolffd@0 78 fullfile=fread(fid,'uchar=>char')';
wolffd@0 79 %Strip LF if free is set
wolffd@0 80 if ~isempty(findstr(options,'free'))
wolffd@0 81 fullfile=strrep(fullfile,char(10),'');
wolffd@0 82 end;
wolffd@0 83 %Find all delimiters
wolffd@0 84 delimpos=[];
wolffd@0 85 for s=1:length(delim)
wolffd@0 86 delimpos=[delimpos find(fullfile==delim(s))];
wolffd@0 87 end
wolffd@0 88 %Find all eol
wolffd@0 89 endpos=find(fullfile==char(10));
wolffd@0 90 endpos=setdiff(endpos,delimpos);
wolffd@0 91 %find all exclusions
wolffd@0 92 xclpos=[];
wolffd@0 93 for s=1:length(exclusions);
wolffd@0 94 xclpos=[xclpos find(fullfile==exclusions(s))];
wolffd@0 95 end
wolffd@0 96 sort(xclpos);
wolffd@0 97 xclpos=[xclpos(1:2:end-1);xclpos(2:2:end)];
wolffd@0 98 %Combine eol and delimiters
wolffd@0 99 jointpos=union(delimpos,endpos);
wolffd@0 100 t=1;
wolffd@0 101 %Remove delim/eol within exclusion pairs
wolffd@0 102 removedelim=[];
wolffd@0 103 for s=1:length(jointpos)
wolffd@0 104 if any((jointpos(s)>xclpos(1,:)) & (jointpos(s)<xclpos(2,:)))
wolffd@0 105 removedelim(t)=jointpos(s);
wolffd@0 106 t=t+1;
wolffd@0 107 end;
wolffd@0 108
wolffd@0 109 end
wolffd@0 110 %and add start point
wolffd@0 111 jointpos=[0 setdiff(jointpos,removedelim)];
wolffd@0 112 i=1;
wolffd@0 113 j=1;
wolffd@0 114 posind=1;
wolffd@0 115 multflag=isempty(findstr(options,'single'));
wolffd@0 116 stringflag=~isempty(findstr(options,'string'));
wolffd@0 117 emptnum=~isempty(findstr(options,'empty2num'));
wolffd@0 118 %Run through
wolffd@0 119 while (posind<(length(jointpos)))
wolffd@0 120 %Get current field
wolffd@0 121 tempstr=fullfile(jointpos(posind)+1:jointpos(posind+1)-1);
wolffd@0 122 %If empty only continue if adjacent delim count.
wolffd@0 123 if ~(isempty(tempstr) & multflag);
wolffd@0 124 %This ij is set
wolffd@0 125 dflag(i,j)=1;
wolffd@0 126 %Convert to num
wolffd@0 127 tempno=str2double([tempstr]);
wolffd@0 128 %If emptystring convert to zero if emptnum set
wolffd@0 129 if (isempty(tempstr) & emptnum)
wolffd@0 130 tempno=0;
wolffd@0 131 end;
wolffd@0 132 %Set dattype to no (or NaN if not a num
wolffd@0 133 dattype(i,j)=tempno;
wolffd@0 134 %If NaN set lc to string else to num if stringflag not set
wolffd@0 135 if (isnan(tempno) | stringflag)
wolffd@0 136 lc{i,j}=tempstr;
wolffd@0 137 else
wolffd@0 138 lc{i,j}=tempno;
wolffd@0 139 end;
wolffd@0 140 %Next j
wolffd@0 141 j=j+1;
wolffd@0 142 end;
wolffd@0 143 %If eol inc i and reset j
wolffd@0 144 if ismember(jointpos(posind+1),endpos)
wolffd@0 145 i=i+1;
wolffd@0 146 j=1;
wolffd@0 147 end;
wolffd@0 148 %Inc to next delim
wolffd@0 149 posind=posind+1;
wolffd@0 150 end;
wolffd@0 151 end;
wolffd@0 152 %Logicalise dflag
wolffd@0 153 dflag=logical(dflag);