所以你在boto3中要求等同于aws s3 ls。这将列出所有顶级文件夹和文件。这是我能得到的最接近的结果;它只列出所有顶级文件夹。这么简单的操作居然这么难。

import boto3

def s3_ls():
  s3 = boto3.resource('s3')
  bucket = s3.Bucket('example-bucket')
  result = bucket.meta.client.list_objects(Bucket=bucket.name,
                                           Delimiter='/')
  for o in result.get('CommonPrefixes'):
    print(o.get('Prefix'))

使用cloudpathlib

cloudpathlib提供了一个方便的包装器，这样您就可以使用简单的pathlib API与AWS S3(以及Azure blob存储、GCS等)进行交互。你可以用pip install "cloudpathlib[s3]"来安装。

像pathlib一样，你可以使用glob或iterdir来列出目录的内容。

下面是一个带有公共AWS S3桶的示例，您可以复制并过去运行该桶。

from cloudpathlib import CloudPath

s3_path = CloudPath("s3://ladi/Images/FEMA_CAP/2020/70349")

# list items with glob
list(
    s3_path.glob("*")
)[:3]
#> [ S3Path('s3://ladi/Images/FEMA_CAP/2020/70349/DSC_0001_5a63d42e-27c6-448a-84f1-bfc632125b8e.jpg'),
#>   S3Path('s3://ladi/Images/FEMA_CAP/2020/70349/DSC_0002_a89f1b79-786f-4dac-9dcc-609fb1a977b1.jpg'),
#>   S3Path('s3://ladi/Images/FEMA_CAP/2020/70349/DSC_0003_02c30af6-911e-4e01-8c24-7644da2b8672.jpg')]

# list items with iterdir
list(
    s3_path.iterdir()
)[:3]
#> [ S3Path('s3://ladi/Images/FEMA_CAP/2020/70349/DSC_0001_5a63d42e-27c6-448a-84f1-bfc632125b8e.jpg'),
#>   S3Path('s3://ladi/Images/FEMA_CAP/2020/70349/DSC_0002_a89f1b79-786f-4dac-9dcc-609fb1a977b1.jpg'),
#>   S3Path('s3://ladi/Images/FEMA_CAP/2020/70349/DSC_0003_02c30af6-911e-4e01-8c24-7644da2b8672.jpg')]

创建于2021-05-21 20:38:47 PDT由reprexlite v0.4.2创建

在上面的注释中对@Hephaeastus的代码进行了少许修改，编写了下面的方法来列出给定路径中的文件夹和对象(文件)。类似s3 ls命令。

from boto3 import session

def s3_ls(profile=None, bucket_name=None, folder_path=None):
    folders=[]
    files=[]
    result=dict()
    bucket_name = bucket_name
    prefix= folder_path
    session = boto3.Session(profile_name=profile)
    s3_conn   = session.client('s3')
    s3_result =  s3_conn.list_objects_v2(Bucket=bucket_name, Delimiter = "/", Prefix=prefix)
    if 'Contents' not in s3_result and 'CommonPrefixes' not in s3_result:
        return []

    if s3_result.get('CommonPrefixes'):
        for folder in s3_result['CommonPrefixes']:
            folders.append(folder.get('Prefix'))

    if s3_result.get('Contents'):
        for key in s3_result['Contents']:
            files.append(key['Key'])

    while s3_result['IsTruncated']:
        continuation_key = s3_result['NextContinuationToken']
        s3_result = s3_conn.list_objects_v2(Bucket=bucket_name, Delimiter="/", ContinuationToken=continuation_key, Prefix=prefix)
        if s3_result.get('CommonPrefixes'):
            for folder in s3_result['CommonPrefixes']:
                folders.append(folder.get('Prefix'))
        if s3_result.get('Contents'):
            for key in s3_result['Contents']:
                files.append(key['Key'])

    if folders:
        result['folders']=sorted(folders)
    if files:
        result['files']=sorted(files)
    return result

这将列出给定路径下的所有对象/文件夹。Folder_path可以默认为None, method将列出桶根目录的即时内容。

也可以这样做:

csv_files = s3.list_objects_v2(s3_bucket_path)
    for obj in csv_files['Contents']:
        key = obj['Key']

为了处理大型键列表(即当目录列表大于1000项时)，我使用以下代码将多个列表中的键值(即文件名)累积起来(感谢上面的阿梅里奥的第一行)。代码是针对python3的:

    from boto3  import client
    bucket_name = "my_bucket"
    prefix      = "my_key/sub_key/lots_o_files"

    s3_conn   = client('s3')  # type: BaseClient  ## again assumes boto.cfg setup, assume AWS S3
    s3_result =  s3_conn.list_objects_v2(Bucket=bucket_name, Prefix=prefix, Delimiter = "/")

    if 'Contents' not in s3_result:
        #print(s3_result)
        return []

    file_list = []
    for key in s3_result['Contents']:
        file_list.append(key['Key'])
    print(f"List count = {len(file_list)}")

    while s3_result['IsTruncated']:
        continuation_key = s3_result['NextContinuationToken']
        s3_result = s3_conn.list_objects_v2(Bucket=bucket_name, Prefix=prefix, Delimiter="/", ContinuationToken=continuation_key)
        for key in s3_result['Contents']:
            file_list.append(key['Key'])
        print(f"List count = {len(file_list)}")
    return file_list

我的s3键实用函数本质上是@Hephaestus的答案的优化版本:

import boto3


s3_paginator = boto3.client('s3').get_paginator('list_objects_v2')


def keys(bucket_name, prefix='/', delimiter='/', start_after=''):
    prefix = prefix.lstrip(delimiter)
    start_after = (start_after or prefix) if prefix.endswith(delimiter) else start_after
    for page in s3_paginator.paginate(Bucket=bucket_name, Prefix=prefix, StartAfter=start_after):
        for content in page.get('Contents', ()):
            yield content['Key']

在我的测试(boto3 1.9.84)中，它比等效的(但更简单)代码要快得多:

import boto3


def keys(bucket_name, prefix='/', delimiter='/'):
    prefix = prefix.lstrip(delimiter)
    bucket = boto3.resource('s3').Bucket(bucket_name)
    return (_.key for _ in bucket.objects.filter(Prefix=prefix))

由于S3保证UTF-8二进制排序结果，因此在第一个函数中添加了start_after优化。