如何在 java 中从头开始读取 arff 文件的属性？

Question

I'm trying to read an arff file from scratch for this assignment. I feel like I'm doing it wrong. So like for example let's say I have this line

@attribute 'habitat' { 'd', 'g', 'l', 'm', 'p', 'u', 'w'}

I want to check if it's an attribute, then take the attribute name, and then add the instances to a list. So far I'm doing this. (st is the current line that i'm reading from the file)

st=st.replaceAll("'", "");
String[] token = st.split(" ");
if(token[0].trim().equals("@attribute")) {
   Attribute a=new Attribute(token[1]);
}

I'm not sure how to read the instances, since I split it according to spaces, the instances were split wrong. I feel like I'm going about reading this file wrong This is my attribute class

public class Attribute {
public String name;
public LinkedList<Instance> instanceList=new LinkedList<Instance>();
}

Any ideas?

Answer 1

Here is how to do it:

your Attribute class

class Attribute {
    private String name;
    private LinkedList<String> instanceList;

    public Attribute(String name, LinkedList<String> instanceList) {
        this.name = name;
        this.instanceList = instanceList;
    }

    public String getName() {
        return name;
    }

    public void setName(String name) {
        this.name = name;
    }

    public LinkedList<String> getInstanceList() {
        return instanceList;
    }

    public void setInstanceList(LinkedList<String> instanceList) {
        this.instanceList = instanceList;
    }

    @Override
    public String toString() {
        return "Attribute{" +
                "name='" + name + '\'' +
                ", instanceList=" + instanceList +
                '}';
    }
}

your main class:

     String str = "@attribute 'habitat' { 'd', 'g', 'l', 'm', 'p', 'u', 'w'}";
     Pattern pattern = Pattern.compile("@attribute '(.*?)' \\{\\s*(.*?)\\s*\\}");
     List<Attribute> list = new ArrayList<>();
     // in case you would like to ignore case sensitivity,
     // you could use this statement:
     // Pattern pattern = Pattern.compile("\\s+", Pattern.CASE_INSENSITIVE);
     Matcher matcher = pattern.matcher(str);
     // check all occurance
     while (matcher.find()) {
                list.add(new Attribute(matcher.group(1),
                        new LinkedList<>(Arrays.asList(matcher.group(2).replaceAll("[',]", "").trim().split("\\s+"))))
         );
            }
     System.out.println(list);

output:

[Attribute{name='habitat', instanceList=[d, g, l, m, p, u, w]}]