Regex for validation of an input file from file

[_max_]

Dear,

I'm quite new to Java so please so be patient with me.

A text file must be validated before being loaded by a program. It contains a list of atom names, distances and affinities for any couple

The format of each line must be exactly as follows:


atom1 atom2 dist12:aff12 atom3 dist13:aff13 atom4 dist14:aff14
atom3 atom5 dist35:aff35
atom5 atom6 dist46:aff56 atom7 dist57:aff57

For instance, a few lines of that file should contain:

Al B 10:2 Cu 15:7 Dy 25:20 C 2:40
Cu I 120:23
I Cl 8:29 Si 123:100

I have tryed to implement a class patternMatch() by using regex from java.util.regex package, but I had no success:

public class patternMatch {

       [B]String regex = "[A-Za-z]+[a-zA-Z0-9]{2,3}+:+[a-zA-Z0-9]{2,3}";[/B]
        
        String input = "Al B 10:2 Cu 15:7 Dy 25:20 C 2:40";
//        String input = "Cu I 120:23";
//        String input = "I Cl 8:29 Si 123:100 ";


        Pattern pattern = Pattern.compile(regex);
        Matcher matcher = pattern.matcher(input);

        while (matcher.find())
        System.out.println(matcher.group());
}

This for getting the atom names, string by string (and this is already a problem, since the file
should be parsed by an I/O stream...) but the result is empty.

Unfortunately I need to validate EACH line from the input file before processing it. No idea on ho to create this complicated REGEX in Java.

Any help for me?

Thanks a lot
M

[JosAH]

Why do people always try to do things with one big ugly moloch of a RE? Why not use the split( ... ) method instead? (read the String class API documenation). The split( ... ) method gives you an array of Strings, say s[], where s[0] is the first atom name s[1] and s[2] give you the second atom name and the dist:affinity pair respectively etc. etc. That pair needs a bit more processing that can also be done by using the split( ... ) method; it'll return an other array, say, da[], where da[0] contains the distance and da[1] contains the affinity. Converting those String values to numbers isn't rocket science either ...

kind regards,

Jos

[_max_]

Dear JosAH,

I am not so confident with Java, as written before.

Can you clarify your suggestion by a piece of code?

Tnx

[JosAH]

Dear JosAH,

I am not so confident with Java, as written before.

Can you clarify your suggestion by a piece of code?

Tnx

Here's an even simpler suggestion:

String line= ...; // a line from your input file
String[] parts= line.spilt("[ :]+"); // split it

Print out every element from that parts array and see for yourself.

kind regards,

Jos

[_max_]

Thank you Jos,

the main question arises from the need to validate each input line, not from reading out its components.

I've tryied the split() method you've suggested and it works, but unfortunately I don't know how to implement the line control. This is because I am to noob, I know :(

Max

[JosAH]

Thank you Jos,

the main question arises from the need to validate each input line, not from reading out its components.

I've tryied the split() method you've suggested and it works, but unfortunately I don't know how to implement the line control. This is because I am to noob, I know :(

Max

Are you saying that you don't want to do anything with the lines and just check whether or not the lines are syntactically valid? If so, a regular expression can do the job easily: compose a RE starting from the small parts: i.e. define the RE for an atom name and define the RE for two integral numbers separated by a colon; glue those together to form the RE for an entire line.
Think of the following details: what characters can an atom name have? How many of them? Integral numbers are (sort of) easy: \d+ does the job. Give it a try. The top level of the RE should look like this: <atom> (<atom> <pair>)+ where <pair> looks like this: \d+:\d+ you do the <atom> part ...

kind regards,

Jos

[_max_]

This is last RE I've tryied without success:

String regex = [B]"(([A-Za-z\\.]+\\s){2}\\d+(\\s[A-Za-z\\.]+\\s\\d+)*[\\n\\r]*)+";[/B]

As test String input = "Aluminium Borum 10:2 Cuprum 15:7 Dys.Prosium 25:20 Carbon 2:40";

No way: the simple class patternMatch as described before prompts:

Aluminium Borum 10

while I am expecting to have the whole line be validated (I mean: printed)