I am processing file inside my method - basically preparing it for the download.

But I decided to split actions for setting headers and reading bytes. As Java does not support method declaration inside methods, I decided to go with class inside method.

Here is the result

public static String downloadFile(final String fileBody, HttpServletResponse response) {
    class FileDownloader {

        private static final String HEADER_KEY = "Content-Disposition";
        private static final String HEADER_VALUE = "attachment; filename=\"%s\"";
        private static final String TEMPLATE_FILE_NAME = "filename.html";

        private void setHeader(HttpServletResponse response) {
            response.setHeader(HEADER_KEY, String.format(HEADER_VALUE, TEMPLATE_FILE_NAME));
        }

        public void processResponse(HttpServletResponse response) throws IOException {
            byte[] bytes = fileBody.getBytes();
            IOUtils.copy(new ByteArrayInputStream(bytes), response.getOutputStream());
            setHeader(response);
        }
    }

    try {
        new FileDownloader().processResponse(response);
        return StringUtils.EMPTY;
    } catch (IOException e) {
        e.printStackTrace();
        return null;
    }
}

But I am not sure whether the idea I've come up with is good or bad. Basically there is no reason for method separation cause they are not doing at lot of work.

I am processing file inside my method - basically preparing it for the download.

But I decided to split actions for setting headers and reading bytes. As Java does not support method declaration inside methods, I decided to go with class inside method.

public static String downloadFile(final String fileBody, HttpServletResponse response) {
    class FileDownloader {

        private static final String HEADER_KEY = "Content-Disposition";
        private static final String HEADER_VALUE = "attachment; filename=\"%s\"";
        private static final String TEMPLATE_FILE_NAME = "filename.html";

        private void setHeader(HttpServletResponse response) {
            response.setHeader(HEADER_KEY, String.format(HEADER_VALUE, TEMPLATE_FILE_NAME));
        }

        public void processResponse(HttpServletResponse response) throws IOException {
            byte[] bytes = fileBody.getBytes();
            IOUtils.copy(new ByteArrayInputStream(bytes), response.getOutputStream());
            setHeader(response);
        }
    }

    try {
        new FileDownloader().processResponse(response);
        return StringUtils.EMPTY;
    } catch (IOException e) {
        e.printStackTrace();
        return null;
    }
}

But I am not sure whether the idea I've come up with is good or bad. Basically there is no reason for method separation because they are not doing a lot of work.