`ContentResolver.openOutputStream(Uri)` does not truncate existing files. If the amount of data written is smaller than the file size, you end up with new data at the beginning of the file followed by old data at the end of the file.