AWSTPEWorkshop-20200320-Lambda-with-Pandas

This workshop will show you how to use Pandas lib in your Lambda function, which implement with Lambda layer.

---

1. Prepare environment

a. Create a cloud9 environment

1. Start to create AWS Cloud9

2. Name environment

3. Configure settings Leave all setting as default and click Next Step

4. Review Click Create environment to create you Cloud9 environment

---

2. Create a Lambda Layer

Create a new Lambda Layer which include Pandas and required python lib.

a. Create an required.txt file

Create an requirement.txt file for pip to install selected lib from file, with File > New File File in navigation bar.

pandas==0.23.4
pytz==2018.7

b. Create a script to run and prepare required resources

Next, create an script called prepare_layer_packages.sh with File > New File File in navigation bar.

#!/bin/bash

export DIR="python"

rm -rf ${DIR} && mkdir -p ${DIR}

docker run --rm -v $(pwd):/foo -w /foo lambci/lambda:build-python3.6 \
    pip install -r requirement.txt --no-deps -t ${DIR}

https://aws.amazon.com/tw/premiumsupport/knowledge-center/lambda-layer-simulated-docker/

c. Package files and download it

Run command as following in your terminal. And then download the my-pandas23-layer.zip in your environment panel and right click it and then Download it.

chmod +x prepare_layer_packages.sh
./prepare_layer_packages.sh
zip -r my-pandas23-layer.zip .

d. Create Lambda layer in AWS

1. Create a new AWS Lambda Layer

2. Input my-pandas23-layer as your Layer name, and then upload your my-pandas23-layer.zip file. In the end, select python 3.6 as runtime.

---

3. Use Lambda layers into functions

Use required Lambda layers into your Lambda functions

a. Creating Lambda function with Lambda Layer

Creating Lambda function and using build-in Lambda layer and customized layer you created.

1. Create a new Lambda function called pandas-sample.

2. Add new Layer to this function

3. Add AWS provided AWSLambda-Python36-SciPy1x

4. Add your layer my-pandas23-layer

b. Testing Lambda function

1. Testing Lambda function working like expected.

import json

import numpy as np
import pandas as pd

def lambda_handler(event, context):
    # TODO implement
    dates = pd.date_range('20181201', periods=6)
    df = pd.DataFrame(np.random.randn(6, 4), index=dates, columns=list('ABCD'))
    print(df)
    return {
        'statusCode': 200,
        'body': json.dumps('Hello from Lambda!')
    }

2. Config a test event

3. Test your Lambda function with Lambda Layer