Reading/writing text file while excluding header

I'm new-ish to MATLAB, and until now I've only used it to perform Matrix Displacement Method analyses for structures classes.

I have large earthquake record files that I would like to read in. They have headers (4 lines) that I would like to omit, and 5 columns of acceleration data (the number of rows depends on the size of the earthquake record). Is there a way to read in this data and then write the array to a new text file? I've spent hours pouring over past forum topics, and I haven't had any luck with my attempts.

The closest I've come is to use importdata(), but the 'data' array returns dimensions that don't seem right (that, and I don't know how to access the elements of a structured array like 'data').

Hints would be much, much appreciated!

[Matt Kindig]

I would use textscan() with the 'HeaderLines' option.

doc textscan

[James]

That's one of the commands that I tried to use. I've had no luck. I was going to post the errors I received but MATLAB was kind enough to crash about 10 minutes ago.

[Matt Kindig]

Can you post the code that you tried?

Also, approximately how many rows do you have?

[James]

Thanks for the responses so far.

My source files look like:

PEER STRONG MOTION DATABASE RECORD. PROCESSING BY PACIFIC ENGINEERING.

FRIULI 09/11/76 1631, BUIA,  UP
ACCELERATION TIME HISTORY IN UNITS OF G. FILTER POINTS: HP=0.4 Hz LP=20.0 Hz
NPTS=  7769, DT= .00500 SEC

   -.6556807E-06  -.6741810E-06  -.6559686E-06  -.6705472E-06  -.6498322E-06
    -.6731312E-06  -.6715338E-06  -.6567325E-06  -.4942821E-06  -.2008082E-06
     .2541901E-06   .7106125E-06   .1093553E-05   .1325778E-05   .1121705E-05
     .1825764E-06  -.1467073E-05  -.3893214E-05  -.7018337E-05  -.1086156E-04
    -.1515249E-04  -.1893733E-04  -.2106391E-04  -.2061759E-04  -.1650684E-04
    -.8531587E-05   .1919265E-05   .1271986E-04   .2007339E-04   .2081788E-04
     .1399449E-04  -.2441929E-06  -.2112721E-04  -.4612119E-04  -.7019579E-04
    -.8699628E-04  -.9050326E-04  -.7702844E-04  -.4508828E-04   .3048132E-05
     .6435106E-04   .1321023E-03   .1944479E-03   .2380203E-03   .2475398E-03

... and can have more than 600 rows.

I have tried the following:

FID=fopen('test_quake.txt');
fprintf(FID,'%f','headerLines',4);
A=fscanf(FID,'%f',[inf]);
fclose(FID);

and

filename = 'test_quake.txt';
delimiterIn = ' ';
headerlinesIn = 4;
A=importdata(filename,delimiterIn,headerlinesIn);

And as I mentioned before, I have tried implementing the textscan() command to no avail.

[Walter Roberson]

In the above there appears to be empty lines in the headers. If those empty lines are present in the file, then you need to include those in the count of header lines.

You should not be using fprintf() for an input file.

Answer 1

[Walter Roberson]

FID=fopen('test_quake.txt');
datacell = textscan(FID, '%f%f%f%f%f', 'HeaderLines', 4, 'CollectOutput', 1);
fclose(FID);
A = datacell{1};

[James]

Thank you for the steady shove in the right direction! However, I'm receiving the error, "??? Error using ==> textscan, Unknown option 'CollectOutput'." Does that make sense to you?

[Walter Roberson]

Which MATLAB version are you using? CollectOutput has been present for a number of versions; http://www.mathworks.com/help/matlab/ref/textscan.html

[James]

I'm using version 7.12.0.635.

[Walter Roberson]

That version definitely supported CollectOutput; see http://www.mathworks.com/help/releases/R2011a/techdoc/ref/textscan.html

Please cross-check the spelling in case of a typo: CollectOutput

Also please use

which -all textscan

just in case you are getting some odd version of the file.

[James]

built-in (/Applications/MATLAB_R2011a.app/toolbox/matlab/iofun/textscan)

Oddly enough, after clearing my command history and workspace, it worked (without changing anything about the original script). I made a few changed to suit my needs (like printing to a text file for the output of the array, and changing the format of the output). Thanks again for all your efforts!

While I have your attention, is it a relatively easy tweak to read the text row by row instead of column-wise?

Thanks again, sincerely.

[Walter Roberson]

FID=fopen('test_quake.txt');
for K = 1 : 4; fgetl(FID); end;   %throw away 4 lines

while true
  [A, count] = fscanf(FID,'%f', 5);
  if count == 0; break; end   %end of file
  .... process the line ...
end

fclose(FID);