Write a binary downloaded file to disk in Java

Question

I have a software that allow to write add-on in javascript files (.js) that allow to use Java function (I don't know if this is common, I never saw java call in javascript file before)

I need to download a binary file from a webserver and write it to the hard drive. I tried the following code:

baseencoder = new org.apache.commons.codec.binary.Base64();
url = new java.net.URL("https://server/file.tgz");

urlConnect = url.openConnection();
urlConnect.setDoInput(true);
urlConnect.setDoOutput(true);
urlConnect.setRequestProperty("authorization","Basic "+ java.lang.String(baseencoder.encodeBase64(java.lang.String( username + ":" + password ).getBytes())));
urlConnect.setRequestProperty("content-type","application/x-www-form-urlencoded");

is = new java.io.DataInputStream(urlConnect.getInputStream());
fstream = new FileWriter("C:\\tmp\\test.tgz");
out = new BufferedWriter(fstream);
while((data = is.read()) != -1){
    out.write(data);
}

out.close();
is.close();

The resulting file is no longer a valid gzip archive. I'm sorry if I did a huge error but I'm not a programmer and don't know Java too much.

Noblenobleman

Answer 1

Don't use a FileWriter - that's trying to convert the data into text.

Just use FileOutputStream.

byte[] buffer = new byte[8 * 1024];

InputStream input = urlConnect.getInputStream();
try {
  OutputStream output = new FileOutputStream(filename);
  try {
    int bytesRead;
    while ((bytesRead = input.read(buffer)) != -1) {
      output.write(buffer, 0, bytesRead);
    }
  } finally {
    output.close();
  }
} finally {
  input.close();
}

Elwood

Also, you'll want to use the correct test condition in your while loop: while((data = is.read()) != null){ ... }. – Epiboly

Thanks using a FileOutputStream works (i still use only a simple read() because it seems that it's not possible to use a byte buffer in javascript as it's not an object. – Noblenobleman

Answer 2

I know this question is already answered, but a simpler approach is to use Apache Commons IO's IOUtils.copy() method, which can fully copy an InputStream to an OutputStream.

Recruitment

Answer 3

DataInputStream is meant for reading Java primitives, not for generic data.

It's also redundant, as urlConnect.getInputStream(); already returns an InputStream, and all InputStreams support read().

is = urlConnect.getInputStream();

P.S. This is assuming is and bis are the same variable. Otherwise, you're reading the wrong stream in the loop.

Tumblebug

Nice Information! Interesting to learn that despite what the naming suggest DataInputStream isn't for generic data..! – Sooksoon

Answer 4

Just read about LimitInputStream sounds like it does exactly what you are doing, buffering the input stream for greater efficiency.

Aster

Answer 5

You can even use NIO FileChannel#transferFrom method.

            URL website = new URL(urlToDownload);

            try (ReadableByteChannel rbc = Channels.newChannel(website.openStream());
                    FileOutputStream fos = new FileOutputStream(filePath);) {
                fos.getChannel().transferFrom(rbc, 0, Long.MAX_VALUE);
            }

Reference link1, link2

Selenium