如何使用S3与sdk将整个“文件夹”复制到另一个路径？

When I do for a single file it works:

当我为一个文件做它时，它工作：

    aws_s3 = AWS::S3.new(S3_CONFIG)
    bucket = aws_s3.buckets[S3_CONFIG["bucket"]]

    object = bucket.objects["user/1/photos/image_1.jpg"]
    new_object = bucket.objects["users/1/photos/image_1.jpg"]
    object.copy_to new_object, {:acl => :public_read}

But I want to move the entire "/photos" folder throws No Such Key. Probably the s3 keys are only the full path for each file. How to do that?

但我想移动整个“/ photos”文件夹抛出No Such Key。可能s3键只是每个文件的完整路径。怎么做？

    aws_s3 = AWS::S3.new(S3_CONFIG)
    bucket = aws_s3.buckets[S3_CONFIG["bucket"]]

    object = bucket.objects["user/1/photos"]
    new_object = bucket.objects["users/1/photos"]
    object.copy_to new_object, {:acl => :public_read}

Thanks!

谢谢！

3 个解决方案

#1

Did it:

做到了：

bucket.objects.with_prefix("user/1/photos").each do |object|
   ...
end

#2

A "folder" is not an object in S3, that is why you can not get it by key, but the folder path is actually a prefix for all the keys of the objects contained by the folder.

“文件夹”不是S3中的对象，这就是您无法通过键获取它的原因，但文件夹路径实际上是文件夹所包含对象的所有键的前缀。

Another important thing, you have to url encode the keys otherwise you may end up with an unknown key error.

另一个重要的事情，你必须对密钥进行url编码，否则你可能会得到一个未知的密钥错误。

require 'aws-sdk'
require 'aws-sdk-s3'
require 'securerandom'
require 'uri'
require "erb"
include ERB::Util

def copy_folder(folder, destination)
    bucket_name = 'your_bucket'
    credentials = Aws::Credentials.new('key', 'secret')
    s3_client = Aws::S3::Client.new(region:'the_region', credentials: credentials)
    enumerate_keys_with_prefix(source).each do |source_object|
       source_key = url_encode(source_object.key)
       destination_key = source_object.key.dup.sub(source, "")
       s3_client.copy_object({bucket: bucket_name, copy_source: bucket_name+'/'+source_key, key: destination+'/'+destination_key, acl: "public-read"})
    end
end

def enumerate_keys_with_prefix(prefix)
    bucket_name = 'your_bucket'
    credentials = Aws::Credentials.new('key', 'secret')
    s3 = Aws::S3::Resource.new(region:'the_region', credentials:credentials)
    return s3.bucket(bucket_name).objects(prefix: prefix)
end

#3

I needed additional code to get this working. Basically chop off the base from the source prefix, then add that to the destination prefix:

我需要额外的代码来实现这一点。基本上从源前缀中删除基数，然后将其添加到目标前缀：

def copy_files_s3(bucket_name, source, destination)
  source_bucket = @s3.buckets[bucket_name]
  source_bucket.objects.with_prefix(source).each do |source_object|
    new_file_name = source_object.key.dup
    new_file_name.slice! source
    new_object = source_bucket.objects["#{destination}#{new_file_name}"]
    source_object.copy_to new_object, {acl: :public_read}
  end
end

#1