Autoload CSVs from s3 into Redshift

redshift_csv_uploader.py

from __future__ import print_function

import json
import urllib
import boto3
import psycopg2
import re

s3 = boto3.client('s3')

def lambda_handler(event, context = {}):
    """
    When you upload `my_csv 2015-01-21.csv` to the s3 bucket your
    lambda is listening to, this copies its contents into `s3.my_csv`
    table in specified Redshift db.
    
    You have to have already made the table in the Redshift db e.g.
    > create schema if not exists s3;
    > create table my_csv (my_column varchar(255));
    
    And to use psycopg2 in a Lambda you have to follow instructions like 
    those found in the following link:
    https://github.com/jkehler/awslambda-psycopg2
    """
    try:
        bucket = event['Records'][0]['s3']['bucket']['name']
        key = urllib.unquote_plus(event['Records'][0]['s3']['object']['key']).decode('utf8')
        table = re.sub(r'(input/|.csv)',"",key).split(" ")[0]
        print("Okay so we're gonna try to copy {0}/{1} to {2}".format(bucket, key, table))

        conn = make_conn()
        print("Connected to Redshift")
        
        query = """
        COPY s3.{0}
        FROM 's3://{1}/{2}'
        CREDENTIALS 'aws_access_key_id=SDFKnotmycredsDSKFJD;aws_secret_access_key=DSFHKD5hjnotmycreds34jjk434343'
        DELIMITER ','
        REMOVEQUOTES
        IGNOREHEADER AS 1;
        """.format(table, bucket, key)
        print("Running query:\n{0}".format(query))
        
        conn.cursor().execute(query)
        print("Query complete...")
        
        conn.commit()
        print("... and committed")
        
        return True
    except Exception as e:
        print(e)
        raise e

def make_conn():
    conn_string = """
    host='name.dfdnotreal343.us-east-1.redshift.amazonaws.com'
    dbname='name'
    user='name'
    password='password'
    port='5439"'
    """
    return psycopg2.connect(conn_string)