tags:

views:

54

answers:

2

I have to parse some tables from an ASCII text file. Here's a partial sample:

QSMDRYCELL   11.00   11.10   11.00   11.00    -.90      11     11000     1.212
RECKITTBEN  192.50  209.00  192.50  201.80    5.21      34      2850     5.707
RUPALIINS   150.00  159.00  150.00  156.25    6.29       4        80      .125
SALAMCRST   164.00  164.75  163.00  163.25    -.45      80      8250    13.505
SINGERBD    779.75  779.75  770.00  773.00    -.89       8        95      .735
SONARBAINS   68.00   69.00   67.50   68.00     .74      11      3050     2.077

The table consists of 1 column of text and 8 columns of floating point numbers. I'd like to capture each column via regex.

I'm pretty new to regular expressions. Here's the faulty regex pattern I came up with:

(\S+)\s+(\s+[\d\.\-]+){8}

But the pattern captures only the first and the last columns. RegexBuddy also emits the following warning:

You repeated the capturing group itself. The group will capture only the last iteration. Put a capturing group around the repeated group to capture all iterations.

I've consulted their help file, but I don't have a clue as to how to solve this.

How can I capture each column separately?

Thanks in advance

+1  A: 

Unfortunately you need to repeat the (…) 8 times to get each column separately.

^(\S+)\s+([-.\d]+)\s+([-.\d]+)\s+([-.\d]+)\s+([-.\d]+)\s+([-.\d]+)\s+([-.\d]+)\s+([-.\d]+)\s+([-.\d]+)$

If code is possible, you can first match those numeric columns as a whole

>>> rx1 = re.compile(r'^(\S+)\s+((?:[-.\d]+\s+){7}[-.\d]+)$', re.M)
>>> allres = rx1.findall(theAsciiText)

then split the columns by spaces

>>> [[p] + q.split() for p, q in allres]
KennyTM
Kenny, thanks for the prompt response! I'm actually using that pattern right now. But I was wondering if there's a better solution using repeating capturing groups.
invarbrass
@invarbrass: Not with repeated capturing groups that I'm aware of. Regexes often work best if you don't try to overdo them with a one-shot.
Owen S.
KennyTM: Thanks! Your solution works - I was doing something similar, albeit a lot less elegantly.
invarbrass
@KennyTM: .NET is special in that it keeps intermediate captures! See Tim's answer and http://stackoverflow.com/questions/3029127/is-there-a-regex-flavor-that-allows-me-to-count-the-number-of-repetitions-matched/
polygenelubricants
+2  A: 

In C# (modified from this example):

string input = "QSMDRYCELL   11.00   11.10   11.00   11.00    -.90      11     11000     1.212";
string pattern = @"^(\S+)\s+(\s+[\d.-]+){8}$";
Match match = Regex.Match(input, pattern, RegexOptions.MultiLine);
if (match.Success) {
   Console.WriteLine("Matched text: {0}", match.Value);
   for (int ctr = 1; ctr < match.Groups.Count; ctr++) {
      Console.WriteLine("   Group {0}:  {1}", ctr, match.Groups[ctr].Value);
      int captureCtr = 0;
      foreach (Capture capture in match.Groups[ctr].Captures) {
         Console.WriteLine("      Capture {0}: {1}", 
                           captureCtr, capture.Value);
         captureCtr++; 
      }
   }
}

Output:

Matched text: QSMDRYCELL   11.00   11.10   11.00   11.00    -.90      11     11000     1.212
...
    Group 2:      1.212
         Capture 0:  11.00
         Capture 1:    11.10
         Capture 2:    11.00
...etc.
Tim Pietzcker
@Tim: Thanks for the heads up. I'm looking into the Group.Captures property.
invarbrass
`Captures` is a neat feature, but it seems like overkill here. Why not just split each line on whitespace? Even if you use the regex to validate the format of the line, it's still less work.
Alan Moore