如何避免压缩上传PDF至S3时在项目目录生成多余PDF文件？

问题根源在于代码里的FileUtils.copyToFile(fileInputStream, pdfFile)这行——它把从classloader读取的PDF流直接写到了本地项目目录里，这是完全多余的操作。我们根本不需要把PDF落地到本地，直接将输入流的数据写入ZipOutputStream就能完成压缩，之后上传S3即可。

修改后的代码 #

public void zipAndUploadToS3() throws IOException {
    // Create a ByteArrayOutputStream to hold the zip data
    ByteArrayOutputStream outputStream = new ByteArrayOutputStream();
    ZipOutputStream zipOutputStream = new ZipOutputStream(outputStream);
    
    for (int i = 1; i <= 3; i++) {
        String pdfFileName = i + ".pdf";
        // Create a new entry in the zip file for the PDF
        ZipEntry entry = new ZipEntry(pdfFileName);
        zipOutputStream.putNextEntry(entry);
        
        InputStream fileInputStream = PdfService.class.getClassLoader().getResourceAsStream(pdfFileName);
        
        if (fileInputStream == null) {
            System.out.println(pdfFileName + " not found");
            zipOutputStream.closeEntry();
            continue;
        }
        
        byte[] buffer = new byte[1024];
        int len;
        while ((len = fileInputStream.read(buffer)) > 0) {
            zipOutputStream.write(buffer, 0, len);
        }
        
        zipOutputStream.closeEntry();
        fileInputStream.close();
    }
    
    zipOutputStream.close();
    
    // Upload the zip file to S3
    byte[] zipData = outputStream.toByteArray();
    ObjectMetadata metadata = new ObjectMetadata();
    metadata.setContentType("application/zip");
    PutObjectRequest putObjectRequest = new PutObjectRequest(bucketName, "my-zip-file.zip", new ByteArrayInputStream(zipData), metadata);
    s3client.putObject(putObjectRequest);
    s3client.setObjectAcl(bucketName, "my-zip-file.zip", CannedAccessControlList.PublicRead);
    String objectUrl = s3client.getUrl(bucketName, "my-zip-file.zip").toString();
    System.out.println(objectUrl);
}

关键改动说明 #

• 移除了FileUtils.copyToFile相关代码，彻底跳过本地PDF文件的生成步骤
• 不再实例化本地File对象，直接用文件名字符串创建ZipEntry
• 处理输入流为空的情况时，新增了关闭当前ZipEntry的操作，避免压缩流程出错

修改后整个压缩过程完全在内存中完成，不会在项目目录生成任何多余文件。

内容的提问来源于stack exchange，提问作者Saurabh Saxena