MATLAB Central Newsreader - read nonuniform text

read nonuniform text

wldf wal — Tue, 28 Oct 2008 09:57:01 -0400

hi,

i have to read nonuniform text. i tried some way to read them but without succes. my text is looking like this:
#cP2008 1 1 0 0 0.00000000 96 ORBIT IGS05 HLM IGS
## 1460 172800.00000000 900.00000000 54466 0.0000000000000
+ 32 G01G02G03G04G05G06G07G08G09G10G11G12G13G14G15G16G17
+ G18G19G20G21G22G23G24G25G26G27G28G29G30G31G32 0 0
+ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
+ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
+ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
++ 2 3 2 2 3 2 3 3 2 3 3 3 3 3 3 3 2
++ 2 2 3 3 2 3 2 3 2 3 3 0 2 2 4 0 0
++ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
++ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
++ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
%c G cc GPS ccc cccc cccc cccc cccc ccccc ccccc ccccc ccccc
%c cc cc ccc ccc cccc cccc cccc cccc ccccc ccccc ccccc ccccc
%f 1.2500000 1.025000000 0.00000000000 0.000000000000000
%f 0.0000000 0.000000000 0.00000000000 0.000000000000000
%i 0 0 0 0 0 0 0 0 0
%i 0 0 0 0 0 0 0 0 0
/* FINAL ORBIT COMBINATION FROM WEIGHTED AVERAGE OF:
/* cod emr esa gfz jpl mit ngs sio
/* REFERENCED TO IGS TIME (IGST) AND TO WEIGHTED MEAN POLE:
/* PCV:IGS05_1461 OL/AL:FES2004 NONE Y ORB:CMB CLK:CMB
* 2008 1 1 0 0 0.00000000
PG01 -14698.277128 -2535.799307 -21763.079031 176.099007 10 9 8 169
PG02 22442.389224 -10272.174156 -9560.136872 162.720385 11 9 10 193
PG03 -13620.463422 8919.821941 20695.070658 166.322185 7 6 8 150
PG04 19958.435008 150.295195 -17787.161153 -44.873145 7 11 8 144
PG05 -2823.115526 -16462.410098 -20939.207089 625.222106 14 9 10 185
PG06 -7143.196104 -25514.299802 1830.117866 170.117599 8 7 8 123
PG07 -10237.510207 -23940.114838 5512.438860 18.039115 11 10 12 127
PG08 14017.015904 7231.477558 21355.061525 -135.440094 12 11 11 183
PG09 12181.805529 -21204.777222 -10751.462189 103.555164 10 8 8 168
PG10 20955.493066 -3772.425757 15961.204995 -191.937552 14 12 13 177
PG11 -7911.583661 23011.271914 -10406.461307 26.858878 11 8 11 164
PG12 3597.925369 -14744.679693 -21709.078627 -353.004481 11 6 7 171
PG13 6302.303623 25466.378831 4141.700761 231.438520 9 9 8 165
PG14 -13901.965954 -14297.974424 -17487.878896 -308.992361 12 11 8 190
PG15 8043.076227 -19262.657551 16459.331387 -70.724287 8 6 7 127
PG16 -23557.523541 435.160643 12502.752577 130.642791 11 12 9 161
PG17 14425.037573 14961.864887 -16604.253188 45.370751 7 10 6 158
PG18 -14740.869594 -17038.887817 14502.939124 -215.440947 6 7 9 155
PG19 -9283.543487 18906.294600 16165.872654 14.687729 9 8 8 157
PG20 -502.268859 15950.765314 -21343.725180 123.445198 10 11 8 192
PG21 -6088.360637 -15011.938405 21125.289387 74.387588 9 8 7 145
PG22 -23711.369538 -11741.131155 2703.641648 206.205087 11 10 10 180
PG23 -1529.915775 24934.711906 -8730.648105 322.616742 11 9 11 162
PG24 7406.306436 -20068.249051 16104.791945 58.635678 10 7 7 164
PG25 167.773732 18905.588920 19085.370779 496.447720 11 7 10 169
PG26 9723.166517 -15936.361423 18196.454530 142.441239 8 6 10 149
PG27 6211.243170 14847.570806 21768.371652 157.876417 12 9 10 169
PG28 22657.886048 13092.931016 5887.875225 -12.977759 10 9 10 184
PG29 3816.647710 -23882.297183 -10941.975317 999999.999999
PG30 -9359.334917 -16911.316266 -18591.680456 58.197189 10 6 7 148
PG31 -20533.999421 2614.616320 -16481.402576 -3.607840 9 7 8 188
PG32 -24787.761236 -6801.202466 -5876.358848 999999.999999
* 2008 1 1 0 15 0.00000000
PG01 -15058.326297 -5000.778117 -21092.964767 176.099823 9 8 8 156
PG02 21733.552608 -9237.464754 -12017.289171 162.723274 10 9 9 182
PG03 -14441.820840 6605.755634 21031.270746 166.327278 7 6 8 147
PG04 18483.506644 1643.505769 -19220.844162 -44.878291 7 10 8 141
PG05 -492.171174 -15995.283710 -21473.031033 625.238575 14 9 9 189
PG06 -6825.928215 -25628.429721 -970.178911 170.104991 8 7 8 122
PG07 -10049.336278 -24451.170524 2761.218462 18.070673 11 10 12 68

first 22 lines is head which i don't need. but lines begining with star char (*) and PGXX i need.

i can't use dlmread (which is great for reading ASCII text) because there are lines with start char (*). and textscan is impractical (for reason a long text - around 3000-4000 lines and i never know number of raws).

there is possibility using non-interactive editor "sed" by i need indempendet on OS.

can somebody help me?

thank you

read nonuniform text

Andres — Tue, 28 Oct 2008 14:58:02 -0400

A suggestion with txt2mat from the file exchange:
.
A = txt2mat('file.txt','ReplaceExpr',{{'*','0'},{'PG','1 '}});
.
The values in the first column of A, 0 or 1, will indicate whether the numbers were taken from a line starting with '*' or 'PG', respectively. Rows will be padded with NaNs.
Hth
Andres

read nonuniform text

wldf wal — Tue, 28 Oct 2008 21:52:01 -0400

"Andres" <rantore@werb.deNoRs> wrote in message <ge799q$ss$1@fred.mathworks.com>...
> A suggestion with txt2mat from the file exchange:
> .
> A = txt2mat('file.txt','ReplaceExpr',{{'*','0'},{'PG','1 '}});
> .
> The values in the first column of A, 0 or 1, will indicate whether the numbers were taken from a line starting with '*' or 'PG', respectively. Rows will be padded with NaNs.
> Hth
> Andres

thank you for you answer :]. I will try your advice (not today but tomorrow :]).

For now I have this code which use function "fgetl()" and then I split line into strings where I convert number in strings to double. I need PGXX number which are converted into number without 'PG' via regular expression.

There is the most important part of my code:

fid = fopen('file.txt');

while ~feof(fid)
  % Read line from file
  newLine = fgetl(fid);
  .
  .
  .
  % Split one string to strings via regular expression
  workLine = regexp(newLine, '([^ \t]*)', 'match');
  .
  .
  .
  % Some lines begining with star (*), so I filter every line and if I catch star (*) I substitude for zero
  if strcmp(newLine(ii),'*')
    SP3data(lineIndex-22,ii) = 0;
  end
  .
  .
  .
  % This part convert "PGXX" number into double number via regular expression
  a = char(workLine(ii));
  aa = regexp(char(a),'([0-9]*)', 'match');
  aaa = char(aa);
  SP3data(lineIndex,ii) = str2double(aaa);
  .
  .
  .
  % This part convert number in string to double
  SP3data(lineIndex-22,ii) = str2double(workLine(ii));
  .
  .
end

read nonuniform text

wldf wal — Tue, 28 Oct 2008 22:04:02 -0400

I forget tell that I have to (!!!) control number of strings/columns in every line. I know the longest line (in my case 9) so I have to add zeros behind numbers (e.g. number number number 0 0 0 0 0 0 ;]).

Or you can create big matrix with 9 columns (contain only zeros) and rewrite every position. What you choose depend on you ;].