views:

2212

answers:

5

How can you read GZIP file in Android located in the "ASSETS" (or resources/raw) folder?

I have tried the following code, but my stream size is always 1.

GZIPInputStream fIn = new GZIPInputStream(mContext.getResources().openRawResource(R.raw.myfilegz)); 
int size = fIn.available();

for some reason the size is always 1. But if Idon't GZIP the file, it works fine.

NOTE: Using Android 1.5

A: 

What happens if you use AssetManager instead of Resources? Example:

InputStream is = mContext.getAssets().open("myfilegz");
GZIPInputStream fIn = new GZIPINputStream(is);

Internally, Resources is just calling AssetManager; I wonder if somewhere along the way it musses things up.

Daniel Lew
A: 

Try looking at the source for Translate from apps-for-android open source project and see if that helps at all.

They use GZIPInputStream on a raw file in their selectRandomWord() function [line 326] (source pasted below)

public void selectRandomWord() {
    BufferedReader fr = null;
    try {
        GZIPInputStream is =
                new GZIPInputStream(getResources().openRawResource(R.raw.dictionary));
snctln
I keep getting a FileNotFoundException which is ridiculus because the Resources know the file is there is R.raw.myfile
Tawani
hmmm, that is an odd problem. It's a shot in the dark but maybe try refreshing your project in eclipse and recompiling as described on http://stackoverflow.com/questions/2322820/update-jar-used-by-android-app-in-eclipse/
snctln
+2  A: 

this is the documented behavior of InflaterInputStream.available:

http://java.sun.com/javase/6/docs/api/java/util/zip/InflaterInputStream.html#available()

Returns 0 after EOF has been reached, otherwise always return 1.

abusing available is a common mistake --- in no case can you assume that it tells you the length of a file (though it sometimes happens to do so, as you've noticed). you want to keep calling read(byte[], int, int) until it returns 0. if you want the length to allocate a byte[] up front, you probably want to create a ByteArrayOutputStream and write to that each time you read, and then get a byte[] from that when you exit the loop. this works for all InputStreams in all cases.

Elliott Hughes
A: 
public class ResLoader {

    /**
     * @param res
     * @throws IOException
     * @throws FileNotFoundException
     * @throws IOException
     */

    static void unpackResources() throws FileNotFoundException, IOException {
        final int BUFFER = 8192;

        android.content.res.Resources t = TestingE3d.mContext.getResources();

        InputStream fis = t.openRawResource(R.raw.resources);
        if (fis == null)
            return;

        ZipInputStream zin = new ZipInputStream(new BufferedInputStream(fis,
                BUFFER));
        ZipEntry entry;
        while ((entry = zin.getNextEntry()) != null) {
            int count;

            FileOutputStream fos = TestingE3d.mContext.openFileOutput(entry
                    .getName(), 0);
            BufferedOutputStream dest = new BufferedOutputStream(fos, BUFFER);

            byte data[] = new byte[BUFFER];

            while ((count = zin.read(data, 0, BUFFER)) != -1) {
                dest.write(data, 0, count);
                // Log.v("NOTAG", "writing "+count + " to "+entry.getName());
            }
            dest.flush();
            dest.close();
        }
        zin.close();

    }

}

R.raw.resources is a zip file - this class will decompress all files in that zip to your local folder. I use this for NDK.

you can access your fils from ndk through: /data/data//files/

package = package where ResLoader resides filename = one of files that is in raw/resources.zip

fazo
A: 

I met the same problem when reading a gz file from assets folder.

It's caused by the file name of the gz file. Just renaming yourfile.gz to other name like yourfile.bin. It seems Android build system would decompress a file automatically if it thought it's a gz.

Cuper Hector