Using regexp in a repetitive manner

Hi,

Given the following string:

[Some text]
type="line" coords="461,461,461,487,461,487,461,489,462,490,462,492,463,493,464,494,465,495,467,495,468,496,470,496,470,496,801,496,801,496,803,496"
[Some text]
type="line" coords="461,487,461,489,462,490,462,492,463,493,496"

How can we obtain the numbers using regexp function of MATLAB? Please note that the length of numbers in the string is variable, therefore, a " universal" expression is needed here.

Thanks in advance,

Answer 1

[Guillaume]

If you want the numbers from all the lines just as a vector and assuming there are no unwanted numbers in any other line then it's simply:

str = sprintf('%s\n%s\n%s\n%s\n', ...
  '[Some text]', ...
  'type="line" coords="461,461,461,487,461,487,461,489,462,490,462,492,463,493,464,494,465,495,467,495,468,496,470,496,470,496,801,496,801,496,803,496"', ...
  'bla bla bla', ...
  'type="line" coords="461,487,461,489,462,490,462,492,463,493,496"');

numbers = str2double(regexp(str, '\d+', 'match'))

if you want something more complicated (such as a cell array of vectors one per line or just the number after coords="), then clarify

[Ive J]

Dear Guillaume ,

Thanks for your answer; however, your assumption is not the case here, because [some text] may contain numeric data, otherwise, it would be easy as you mentioned in your response.

[Guillaume]

So, what do you want? Just the number after any 'coords = "'? Only when the 'type' is 'line'? What are the restrictions?

[Guillaume]

If the numbers have to be enclosed within the coords quotes, then you need a minimum of two regexp:

coordsegments = regexp(str, '(?<=coords=")[^"]*', 'match');
coords = regexp(coordsegments', '\d+', 'match')
coordvalues = cellfun(@str2double, coords, 'UniformOutput', false)

[Ive J]

Yes, that is exactly what I was looking for.

Deeply appreciate your time.

Answer 2

[Thorsten]

line = 'type="line" coords="461,461,461,487,461,487,461,489,462,490,462,492,463,493,464,494,465,495,467,495,468,496,470,496,470,496,801,496,801,496,803,496"'
data = sscanf(strrep(line, 'type="line" coords="', ''), '%d,');

line = 'type="line" coords="461,487,461,489,462,490,462,492,463,493,496"';
data = sscanf(strrep(line, 'type="line" coords="', ''), '%d,');

[Ive J]

Dear Throsten,

Thanks for your response. However, your code returns an empty answer (assuming "line" in your code contains the given string). Please note that [Some text] in the given string may be problematic in this approach, since %d correspond to "any" number in the string, and that's not the case.

[Thorsten]

Dear Ive,

I changed the code. If you can process the text line by line, you can extract the data. It works for different number of coords, and ignores all lines that do not start with 'type="line" coords="'.

[Ive J]

Unfortunately, it is not possible to distinguish among lines. Nevertheless, I deeply appreciate your time and help.