How to read a file from jar in Java?

Question

I want to read an XML file that is located inside one of the jars included in my class path. How can I read any file which is included in the jar?<

Answer 1

Ah, this is one of my favorite subjects. There are essentially two ways you can load a resource through the classpath:

Class.getResourceAsStream(resource)

and

ClassLoader.getResourceAsStream(resource)

(there are other ways which involve getting a URL for the resource in a similar fashion, then opening a connection to it, but these are the two direct ways).

The first method actually delegates to the second, after mangling the resource name. There are essentially two kinds of resource names: absolute (e.g. "/path/to/resource/resource") and relative (e.g. "resource"). Absolute paths start with "/".

Here's an example which should illustrate. Consider a class com.example.A. Consider two resources, one located at /com/example/nested, the other at /top, in the classpath. The following program shows nine possible ways to access the two resources:

package com.example;

public class A {

    public static void main(String args[]) {

        // Class.getResourceAsStream
        Object resource = A.class.getResourceAsStream("nested");
        System.out.println("1: A.class nested=" + resource);

        resource = A.class.getResourceAsStream("/com/example/nested");
        System.out.println("2: A.class /com/example/nested=" + resource);

        resource = A.class.getResourceAsStream("top");
        System.out.println("3: A.class top=" + resource);

        resource = A.class.getResourceAsStream("/top");
        System.out.println("4: A.class /top=" + resource);

        // ClassLoader.getResourceAsStream
        ClassLoader cl = A.class.getClassLoader();
        resource = cl.getResourceAsStream("nested");        
        System.out.println("5: cl nested=" + resource);

        resource = cl.getResourceAsStream("/com/example/nested");
        System.out.println("6: cl /com/example/nested=" + resource);
        resource = cl.getResourceAsStream("com/example/nested");
        System.out.println("7: cl com/example/nested=" + resource);

        resource = cl.getResourceAsStream("top");
        System.out.println("8: cl top=" + resource);

        resource = cl.getResourceAsStream("/top");
        System.out.println("9: cl /top=" + resource);
    }

}

The output from the program is:

1: A.class nested=java.io.BufferedInputStream@19821f
2: A.class /com/example/nested=java.io.BufferedInputStream@addbf1
3: A.class top=null
4: A.class /top=java.io.BufferedInputStream@42e816
5: cl nested=null
6: cl /com/example/nested=null
7: cl com/example/nested=java.io.BufferedInputStream@9304b1
8: cl top=java.io.BufferedInputStream@190d11
9: cl /top=null

Mostly things do what you'd expect. Case-3 fails because class relative resolving is with respect to the Class, so "top" means "/com/example/top", but "/top" means what it says.

Case-5 fails because classloader relative resolving is with respect to the classloader. But, unexpectedly Case-6 also fails: one might expect "/com/example/nested" to resolve properly. To access a nested resource through the classloader you need to use Case-7, i.e. the nested path is relative to the root of the classloader. Likewise Case-9 fails, but Case-8 passes.

Remember: for java.lang.Class, getResourceAsStream() does delegate to the classloader:

     public InputStream getResourceAsStream(String name) {
        name = resolveName(name);
        ClassLoader cl = getClassLoader0();
        if (cl==null) {
            // A system class.
            return ClassLoader.getSystemResourceAsStream(name);
        }
        return cl.getResourceAsStream(name);
    }

so it is the behavior of resolveName() that is important.

Finally, since it is the behavior of the classloader that loaded the class that essentially controls getResourceAsStream(), and the classloader is often a custom loader, then the resource-loading rules may be even more complex. e.g. for Web-Applications, load from WEB-INF/classes or WEB-INF/lib in the context of the web application, but not from other web-applications which are isolated. Also, well-behaved classloaders delegate to parents, so that duplicateed resources in the classpath may not be accessible using this mechanism.