Question

I have a folder containing 100000 files, and need to get 1000 files from this folder through random sampling. Are there any sample functions that I can use to sample from folder? In addition, how to copy the sampled files to another folder?

Answer 1

Random selection could follow along the following lines

File files[] = new File("/path/to/files").listFiles();
Map<Integer, File> selection = new HashMap<Integer, File>(1000);
while (selection.size() < 1000) {
    int value = (int)Math.round(Math.random() * files.length);
    if (!selection.containsKey(value)) {
        selection.put(value, files[value]);
    }
}
for (File file : selection.values()) {
    System.out.println(file);
}

Essentially, you need to grab a list of the available files and the randomly pick through the list until you have enough of a sample. Check out java.io.File

There are plenty of examples of file copying over the net (and SO). If you’re really stuck you could have a look the IO Trail or Apache Commons IO which I believe has a utility class capable of coping files

UPDATED

As suggested by Andrew, you could simply shuffle the file list and pull the first 1000 elements…

File files[] = new File("/path/to/files").listFiles();
List<File> selection = null;
List<File> fileList = new ArrayList<File>(Arrays.asList(files));
Collections.shuffle(fileList);
selection = fileList.subList(0, Math.min(1000, fileList.size()));

for (File file : selection) {
    System.out.println(file);
}