sscanf how to skip lines - not textscan!!!

Is there a way to skip the first X lines in a text file when reading a text file using sscanf? I know this can be done using textscan but I've really struggled to get to grips using it for my particular situation and was wondering if there's an easy way to do it for sscanf.

My text file looks like:

bla bla bla bla

{

  position 99.35 56.73 116.41
  gamma_angle 90
  weight 0.800000
  collimator_setup  16 16 16 8 16 16 16 8

}

With the latter part in '{}' repeated many times. My current code looks like this:

str = fileread('filename.txt');
mat = sscanf(str,'{\r\r  position %f %f %f\r  gamma_angle %f\r  weight %f\r  collimator_setup  %f %f %f %f %f %f %f %f\r\r}\r\r\r\r', [13,Inf]).';

The code works perfectly when the 'bla bla bla bla' is removed and returns a table of 13 collumns with the positions, angles etc...

However when any text is in front of the first '{', the code no longer works.

Thanks in advance.

[dpb]

Why not just remove the empty records first?

[Hannah Taylor]

sorry im not exactly sure what you're asking? If you're asking why not just remove 'bla bla bla bla' manually, its because I have thousands of text files so it would take ages. If you're saying I could write it into my script to remove it, please may you say / point me in the direction of how to do this. Thanks!

[dpb]

Give us a sample file...only needs contain the pertinent parts...it's hard to guess without hard data to look at.

Of course, it also raises the question of if you can do what want/need w/ textscan, why not use it instead of trying to beat sscanf into submission?

[Mathieu NOE]

hello

simple work around built on your code. They may be smarter way to do it, but ...

str = fileread('Document1.txt');
ind1 = findstr(str,'{');
ind2 = findstr(str,'}');
str1 = str(ind1:ind2);
mat = sscanf(str1,'{\r\r  position %f %f %f\r  gamma_angle %f\r  weight %f\r  collimator_setup  %f %f %f %f %f %f %f %f\r\r}\r\r\r\r', [13,Inf]).';

Answer 1

[Walter Roberson]

You cannot do that with sscanf().

I recommend that instead you use regexp() with named tokens and the 'names' option . That will return a struct array in which the entries are character vectors. You can then do things like

gamma_angles = str2double({Parts.gamma_angle});

[Walter Roberson]

Also you can potentially use textscan with Header lines option. You can pass a character vector to textscan instead of a file identifier.

Answer 2

[Stephen23]

"Is there a way to skip the first X lines in a text file when reading a text file using sscanf?"

No, because sscanf parses string/character vectors, not files. To efficiently parse files use fscanf:

https://www.mathworks.com/help/matlab/ref/fscanf.html

"and was wondering if there's an easy way to do it for sscanf."

You can with one extra fscanf call to read (and discard) the file data up until the first "{" character (this relies on one condition: the "{" character must NOT exist anywhere in the"bla bla bla" text, but any other characters can).

fid = fopen('temp1.txt','rt');
fscanf(fid,'%*[^{]') % <- this answers your question.
fmt = '{\n\n  position%f%f%f\n  gamma_angle%f\n  weight%f\n  collimator_setup%f%f%f%f%f%f%f%f\n\n}\n';
mat = fscanf(fid, fmt, [13,Inf]).'
fclose(fid);

I created a sample file (attached, because you did not supply one) and tested the code, it imported everything correctly:

mat =
    99.35000    56.73000   116.41000    90.00000     0.80000    16.00000    16.00000    16.00000     8.00000    16.00000    16.00000    16.00000     8.00000
    99.36000    56.74000   116.42000    91.00000     0.80000    17.00000    17.00000    17.00000     9.00000    17.00000    15.00000    15.00000     7.00000