AWS Lambda：如何在不保存文件的情况下流式传输文件

Question

I want to run a node function in AWS lambda, and let users upload and analyze their files using a web browser. Here is the aforementioned function:

analyzer.analyze(fs.createReadStream('./path/to/local/file'),function(err,analysis){
    console.log(JSON.stringify(analysis));
});

As you can see, the first argument of analyzer.analyze method takes a stream. How can I create that stream without actually saving the file when running it on AWS lambda?

Is there a way to let lambda directly read the file stream from the request, or maybe read the file in memory and then create a stream from there?

Answer 1

Are you fine with storing the file in S3? If yes, you can store the file in S3 and configure the lambda function to trigger upon file creation in that S3 bucket.

Answer 2

The createReadStream requires a file. The lambda environment has read-only partitions, except the partition /tmp . If you just have the createReadStream as option, I'd suggest you to save your file in the /tmp folder instead to upload it in a S3 Bucket.