How to parse a table (in a text file) into matlab cell array where we have some missing values?

Hello all.

How to parse this table into a cell string (or even a vector for each variable)? Each column represent a variable (we have 5 variables). Please note that we need to ignor "- Psat " also most importantnly some columns have missing values. I would like to just put an empty space (or -999) in those places where we have no value. Also there are some empty spaces at the begining of the first column.

This table is generated by a software and lots of such files are generated. Please note that different tables have the missing values in different columns. However, the number of columns is always 5.

 ------------------- ------------ ------------ ------------ ------------ 
    2000.000               0.9852                   45.7751              
    1800.000               0.9878                   45.6544              
    1600.000               0.9905                   45.5293              
    1400.000               0.9933                   45.3995              
    1200.000               0.9963                   45.2648              
    1000.000               0.9994                   45.1246              
     961.666 - Psat        1.0000                   45.0971       3.6734 
     800.000               1.1361       0.0553      45.6508       3.0676 
     600.000               1.4394       0.1258      46.3887       2.3402 
     400.000               2.1396       0.2041      47.2512       1.6256 
     200.000               4.6874       0.3085      48.4664       0.8928 
 ------------------- ------------ ------------ ------------ ------------ 

[Guillaume]

That's not a very clever format. In particular, the text in the middle of the numbers is not helping.

It looks like it's fixed width. Does the width of the column changes from file to file? In particular, does the width of the 1st column depend on the length of the text (if present)?

[RR RR]

I guess the length of columns are 20,12,12,12,12

Answer 1

[Guillaume]

Your file appears to have fixed width fields. the easiest way to import such files is with FixedWidthImportOptions, e.g.:

opts = opts = fixedWidthImportOptions('NumVariables', 5, ...
                                      'VariableNames', {'Pressure', 'something', 'somethingelse', 'andanother', 'xyz'}, ...
                                      'VariableWidths', [21, 13, 13, 13, 13], ...
                                      'VariableTypes', repelem({'double'}, 5), ...
                                      'CommentStyle', '-');
data = readtable('new1.txt', opts)

Note that I included the separating space as part of the field width hence they're 1 higher than you specified.

This works well except for the value that includes '- Psat' since that's no longer convertible to a number. Perhaps the best way to cope with that is to add another variable for that column:

opts = fixedWidthImportOptions('NumVariables', 6, ...
                               'VariableNames', {'Pressure', 'Comment', 'something', 'somethingelse', 'andanother', 'xyz'}, ...
                               'VariableWidths', [13, 8, 13, 13, 13, 13], ...
                               'VariableTypes', [{'double', 'string'}, repelem({'double'}, 4)], ...
                               'CommentStyle', '-');
data = readtable('new1.txt', opts)

Of course, all of this only works if all your files have the column width.