Extracting Time Data from Text File

In short, I've been trying to put together a script that scans a .txt file and pulls number values from it. The context is Rubik's Cube times; I have a program that produces .txt files from sessions, and the contents include somee basic statistics at the top, a list of times, and the moves used to scramble the cube for each time. I've included an example of the beginning of one of these files below:

Average: 10.05

Best: 9.34

Worst: 14.26

Mean: 10.75

Standard Deviation: 1.79

1: 9.91 D2 L2 B2 L2 F2 D L2 B2 D' U L' U' L D' L F' R' B' U B2

2: 9.86 R2 U2 B2 F2 D L2 D' U' R2 U' B2 R U F' U' B' D B2 U R' D2 F2

3: (14.26) D2 L2 B' L2 B2 D2 L2 F2 D2 F R' D B R2 B' D' R2 B2 F L

4: (9.34) B F U2 B R2 B' U2 F' D2 R' D' B2 D' B' U2 R' B' D' B'

5: 10.39 B F' U2 B2 L2 U2 R2 B' R2 B U B2 R2 D2 R2 B' L B R D2

My goal is to pull the times from these files to run my own stats on them, but there are a lot of nuances that have made this difficult!

The biggest issues I've had:

1) As you can see, some of the times contain 3 sig figs, and some contain 4. I haven't found a clean way to pull both from the file at the same time.

2) I haven't found a way to ignore the statistics at the top (if this is too much for a script, it would be easy to just delete that section of the text file before I run the script.

What I have so far isn't pretty, but so far this has successfully created an array of the 4-digit times for me:

fileID = fopen("txt Files/text-4AB30AA43EF2-1.txt")

raw = fscanf(fileID,"%c")

fclose(fileID)

num = '[0123456789][0123456789].[0123456789][0123456789]';

out = regexp(raw,num,'match');

gen1 = str2double(out);

list = gen1'

Any help would be greatly appreciated, thank you!

Answer 1

[Stephen23]

Simpler and more efficient:

str = fileread('test_1.txt');
tkn = regexp(str,'^\s*(\d+):\D+(\d+\.?\d*)\W+([^\n\r]+)', 'tokens', 'lineanchors');
tkn = vertcat(tkn{:})

tkn = 5×3 cell array
    {'1'}    {'9.91' }    {'D2 L2 B2 L2 F2 D L2 B2 D' U L' U' L D' L F' R' B' U B2'      }
    {'2'}    {'9.86' }    {'R2 U2 B2 F2 D L2 D' U' R2 U' B2 R U F' U' B' D B2 U R' D2 F2'}
    {'3'}    {'14.26'}    {'D2 L2 B' L2 B2 D2 L2 F2 D2 F R' D B R2 B' D' R2 B2 F L'      }
    {'4'}    {'9.34' }    {'B F U2 B R2 B' U2 F' D2 R' D' B2 D' B' U2 R' B' D' B''       }
    {'5'}    {'10.39'}    {'B F' U2 B2 L2 U2 R2 B' R2 B U B2 R2 D2 R2 B' L B R D2'       }

mat = str2double(tkn(:,1:2))

mat = 5×2

1.0000 9.9100 2.0000 9.8600 3.0000 14.2600 4.0000 9.3400 5.0000 10.3900

Optional extras:

% tkn(:,1:2) = num2cell(mat); % optional, but it is much better to store numeric
%                               data in a numeric array, so probably best avoided.
spl = regexp(tkn(:,3),'\w+','match')

spl = 5×1 cell array
    {1×20 cell}
    {1×22 cell}
    {1×20 cell}
    {1×19 cell}
    {1×20 cell}

spl{:}

ans = 1×20 cell array
    {'D2'}    {'L2'}    {'B2'}    {'L2'}    {'F2'}    {'D'}    {'L2'}    {'B2'}    {'D'}    {'U'}    {'L'}    {'U'}    {'L'}    {'D'}    {'L'}    {'F'}    {'R'}    {'B'}    {'U'}    {'B2'}

ans = 1×22 cell array
    {'R2'}    {'U2'}    {'B2'}    {'F2'}    {'D'}    {'L2'}    {'D'}    {'U'}    {'R2'}    {'U'}    {'B2'}    {'R'}    {'U'}    {'F'}    {'U'}    {'B'}    {'D'}    {'B2'}    {'U'}    {'R'}    {'D2'}    {'F2'}

ans = 1×20 cell array
    {'D2'}    {'L2'}    {'B'}    {'L2'}    {'B2'}    {'D2'}    {'L2'}    {'F2'}    {'D2'}    {'F'}    {'R'}    {'D'}    {'B'}    {'R2'}    {'B'}    {'D'}    {'R2'}    {'B2'}    {'F'}    {'L'}

ans = 1×19 cell array
    {'B'}    {'F'}    {'U2'}    {'B'}    {'R2'}    {'B'}    {'U2'}    {'F'}    {'D2'}    {'R'}    {'D'}    {'B2'}    {'D'}    {'B'}    {'U2'}    {'R'}    {'B'}    {'D'}    {'B'}

ans = 1×20 cell array
    {'B'}    {'F'}    {'U2'}    {'B2'}    {'L2'}    {'U2'}    {'R2'}    {'B'}    {'R2'}    {'B'}    {'U'}    {'B2'}    {'R2'}    {'D2'}    {'R2'}    {'B'}    {'L'}    {'B'}    {'R'}    {'D2'}

[Stephen23]

A breakdown of that regular expression, the three token groups collect the desired data:

^\s*(\d+):\D+(\d+\.?\d*)\W+([^\n\r]+)
^                                     match the start of the line
 \s*                                  match zero or more whitespace (optional)
    (   )                             1st token group
     \d+                              match one or more digits
         :                            match literal colon
          \D+                         match one or more non-digits
             (         )              2nd token group
              \d+\.?\d*               match digits with optional decimal point
                        \W+           match non-word characters
                           (        ) 3rd token group
                            [^\n\r]+  match one or more non-newlines

[Simon Montrose]

Wow, thank you so much Stephen! Definitely a more efficient route, and I really appreciate you splitting up the regular expression the way you did -- I'm still new to conversion characters and how they interact, you've passively taught me a lot here.

[Stephen23]

@Simon Montrose: if my explanation helped you please vote for my answer :)

[Star Strider]

Definitely worth voting for!

+1

[Simon Montrose]

@Stephen Of course, I didn't know I could switch them my apologies! Still getting acquainted with the forums, but I've implemented your suggestions and have the script completely up and running :) Cheers!