如何在 Python 中使用 Pandas 从 s3 存储桶中读取 csv 文件

Question

我正在尝试使用以下代码将位于 AWS S3 存储桶中的 CSV 文件作为 pandas 数据帧读入内存：

import pandas as pd
import boto

data = pd.read_csv('s3:/example_bucket.s3-website-ap-southeast-2.amazonaws.com/data_1.csv')

为了提供完全访问权限，我在 S3 存储桶上设置了存储桶策略，如下所示：

{
"Version": "2012-10-17",
"Id": "statement1",
"Statement": [
    {
        "Sid": "statement1",
        "Effect": "Allow",
        "Principal": "*",
        "Action": "s3:*",
        "Resource": "arn:aws:s3:::example_bucket"
    }
]

}

不幸的是，我仍然在 python 中收到以下错误：

boto.exception.S3ResponseError: S3ResponseError: 405 Method Not Allowed

想知道是否有人可以帮助解释如何在 AWS S3 中正确设置权限或正确配置 pandas 以导入文件。谢谢！

Answer 1

使用熊猫 0.20.3

import os
import boto3
import pandas as pd
import sys

if sys.version_info[0] < 3: 
    from StringIO import StringIO # Python 2.x
else:
    from io import StringIO # Python 3.x

# get your credentials from environment variables
aws_id = os.environ['AWS_ID']
aws_secret = os.environ['AWS_SECRET']

client = boto3.client('s3', aws_access_key_id=aws_id,
        aws_secret_access_key=aws_secret)

bucket_name = 'my_bucket'

object_key = 'my_file.csv'
csv_obj = client.get_object(Bucket=bucket_name, Key=object_key)
body = csv_obj['Body']
csv_string = body.read().decode('utf-8')

df = pd.read_csv(StringIO(csv_string))

Answer 2

基于 this answer 建议使用 smart_open 从 S3 读取，这就是我在 Pandas 中使用它的方式：

import os
import pandas as pd
from smart_open import smart_open

aws_key = os.environ['AWS_ACCESS_KEY']
aws_secret = os.environ['AWS_SECRET_ACCESS_KEY']

bucket_name = 'my_bucket'
object_key = 'my_file.csv'

path = 's3://{}:{}@{}/{}'.format(aws_key, aws_secret, bucket_name, object_key)

df = pd.read_csv(smart_open(path))

Answer 3

你不需要pandas..你可以使用python的默认csv库

def read_file(bucket_name,region, remote_file_name, aws_access_key_id, aws_secret_access_key):
    # reads a csv from AWS

    # first you stablish connection with your passwords and region id

    conn = boto.s3.connect_to_region(
        region,
        aws_access_key_id=aws_access_key_id,
        aws_secret_access_key=aws_secret_access_key)

    # next you obtain the key of the csv you want to read
    # you will need the bucket name and the csv file name

    bucket = conn.get_bucket(bucket_name, validate=False)
    key = Key(bucket)
    key.key = remote_file_name
    data = key.get_contents_as_string()
    key.close()

    # you store it into a string, therefore you will need to split it
    # usually the split characters are '\r\n' if not just read the file normally 
    # and find out what they are 

    reader = csv.reader(data.split('\r\n'))
    data = []
    header = next(reader)
    for row in reader:
        data.append(row)

    return data

希望它能解决你的问题， 祝你好运！ :)

Answer 4

我最终意识到，您还需要为存储桶中的每个单独对象设置权限，以便使用以下代码提取它：

from boto.s3.key import Key
k = Key(bucket)
k.key = 'data_1.csv'
k.set_canned_acl('public-read')

而且我还得修改pd.read_csv命令中的bucket地址，如下：

data = pd.read_csv('https://s3-ap-southeast-2.amazonaws.com/example_bucket/data_1.csv')