Importing only specific data from a text file

I have a text file with data such as the following:

M13=1 M14=5000 M15=0 M16=-12.5 M17=100 etc

the problem is I only want to import some, not all of this data, and save them to an mfile.

How can I tell matlab what data to include and what to exclude?

[Michael]

the data are all on individual lines in the text file

[Geoff Hayes]

What is your criteria for importing/picking one line of data and not another?

[Michael]

Only certain ones that I need. For instance, I don't need M14, but I do want M15 and M16, and I don't want M17, etc.

[Michael]

there are also character strings within the data that I want to ignore also

[José-Luis]

I am afraid you are not answering the question. In order to import your data efficiently you need to define the pattern for importing it, e.g. "pick one line then skip two lines until you reach the end of the file, beware that there are three characters at the beginning of the line".

Alternatively, you need to hardcode the data you want: "I want M13 and M42".

Otherwise it's not going to automagically happen, although functions such as importdata work small miracles these days.

[Michael]

It would be best to hardcode the data to specify M13, M42, etc.

[Geoff Hayes]

In that case you are going to have to read in each line, and do a strfind within that line for all possible allowed values of M13,M42, etc. How many of these allowed strings are there? Is there no pattern? Do any of the lines in your text file repeat i.e.

 M13=1
 M14=5000
 etc.
 M13=2  % repeated prefix of M13

[Michael]

there is no pattern and no repeats. There's several thousand lines of variables, I only need the values of about 20 of them. Some start with P some start with M, some start with Q, etc. I'm only interested in about 20 that start with M, if that helps.

[dpb]

I'd recommend to forget about only reading the lines wanted in place of scarfing up the whole file as string array and then doing either regexp or regular string-matching to find the rows desired. Then, having that array index, delete the remaining lines and use textscan or whatever other useful i/o functions (again, perhaps including regular expressions) to convert.

For only "several thousand" unless you know the ones desired are within the first few lines it's probably just as fast to read the whole file as it is to process line-by-line.

To get the file into memory

file=textread('yourfile.dat', '%s', 'delimiter', '\n', ...
              'whitespace', '');

If you were able to a priori know a range of lines within the file, you could limit the amount imported by adding 'headerlines' and a repeat count on the format to load only a given section of known size. But, that's a purely mechanical count so would have to know something about the file structure independent of content.