C# - Regex - Matching file names according to a specific naming pattern

Question

I have an application which needs to find and then process files which follow a very specific naming convention as follows.

IABC_12345-0_YYYYMMDD_YYYYMMDD_HHMMSS.zip

I cant see any easy way of doing this using a search pattern so Im assuming Ill have to do something like this after I have generated a list of files using a simpler wildcard pattern.

RegEx re = new RegEx("blah");

foreach(FileInfo fi in Directory.GetFiles(path, "I*.zip"))
{
    if(re.IsMatch(fi.Name))
       //blah blah blah
}

Is this the best way of doing this, and if so, how would I form a regular expression to match this file format?

Answer 1

    string pattern = @"I[A-Z]{3}_\d{5}-\d_\d{8}_\d{8}_\d{6}\.zip";
    var matches = Directory.GetFiles(@"c:\temp")
        .Where(path => Regex.Match(path, pattern).Success);

    foreach (string file in matches)
        Console.WriteLine(file); // do something

Answer 2

It depends on how specific you want to match those names. Is this specific enough:

I[A-Z]{3}_\d{5}-\d_\d{8}_\d{8}_\d{6}\.zip

?

Explanation:

I             // match an 'I'
[A-Z]{3}      // followed by three upper case letters
_             // followed by an underscore
\d{5}         // followed by five digits
-             // followed by a hyphen
\d            // followed by a single digit
_             // followed by an underscore
\d{8}         // followed by eight digits
_             // followed by an underscore
\d{8}         // followed by eight digits
_             // followed by an underscore
\d{6}         // followed by six digits
\.zip         // followed by '.zip'

But, if you have files whose names contain invalid dates or times, it cannot practically be done with regex alone, especially if your DATE_DATE part specifies a date range. You will have to match all file names like I (and others) have shown you, and then perform some "regular" programming logic to filter out the invalid ones.

Answer 3

For a simple regular expression that will also match invalid time specifications (ie. hours=73 etc.), you could use something like this:

^I[A-Z]{3}_\d{5}-\d_\d{8}_\d{8}_\d{6}\.zip$