saving csv file to s3 using boto3

Question

I am trying to write and save a CSV file to a specific folder in s3 (exist). this is my code:

from io import BytesIO
import pandas as pd
import boto3
s3 = boto3.resource('s3')

d = {'col1': [1, 2], 'col2': [3, 4]}
df = pd.DataFrame(data=d)

csv_buffer = BytesIO()

bucket = 'bucketName/folder/'
filename = "test3.csv"
df.to_csv(csv_buffer)
content = csv_buffer.getvalue()

def to_s3(bucket,filename,content):
  s3.Object(bucket,filename).put(Body=content)

to_s3(bucket,filename,content)

this is the error that I get:

Invalid bucket name "bucketName/folder/": Bucket name must match the regex "^[a-zA-Z0-9.\-_]{1,255}$"

I also tried :

bucket = bucketName/folder

and:

bucket = bucketName
key = folder/
s3.Object(bucket,key,filename).put(Body=content)

Any suggestions?

Answer 1

Saving into s3 buckets can be also done with upload_file with an existing .csv file:

import boto3
s3 = boto3.resource('s3')

bucket = 'bucket_name'
filename = 'file_name.csv'
s3.meta.client.upload_file(Filename = filename, Bucket= bucket, Key = filename)

Answer 2

This should work

def to_s3(bucket,filename, content):
    client = boto3.client('s3')
    k = "folder/subfolder"+filename
    client.put_object(Bucket=bucket, Key=k, Body=content)

Answer 3

This should work:

bucket = bucketName
key = f"{folder}/{filename}"
csv_buffer=StringIO()
df.to_csv(csv_buffer)
content = csv_buffer.getvalue()
s3.put_object(Bucket=bucket, Body=content,Key=key)

AWS bucket names are not allowed to have slashes ("/"), which should be part of Key. AWS uses slashes to show "virtual" folders in the dashboard. Since csv is a text file I'm using StringIO instead of BytesIO

Answer 4

This works for me.

import os
import pandas as pd
import boto3
from io import StringIO

from dotenv import load_dotenv
load_dotenv("/path/to/.env", override=True)


def df_to_s3(df, bucket, key):
    # Create a session
    session = boto3.session.Session(profile_name=os.environ.get("AWS_SECRETS_PROFILE_NAME"))
    aws_s3_client = session.client(
        service_name="s3",
        region_name=os.environ.get("AWS_SECRETS_REGION_NAME"),
    )

    # Create a CSV string from the DataFrame
    csv_buffer = StringIO()
    df.to_csv(csv_buffer, index=False)

    # Put the CSV string to S3
    aws_s3_client.put_object(
        Body=csv_buffer.getvalue(),
        Bucket=bucket,
        Key=key
    )
    print(f'Successfully put DataFrame to {bucket}/{key}')

df = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})
df_to_s3(df, 'bucketName/folder/', 'test3.csv')