Last active
September 27, 2021 12:48
-
-
Save t3ndai/3d79a4432ea22786d9d4c53d3d28a4a1 to your computer and use it in GitHub Desktop.
python etl
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
if __name__ == "__main__": | |
import pandas as pd | |
from english_category_name import * | |
from dask.distributed import Client, LocalCluster | |
cluster = LocalCluster() | |
client = Client(cluster) | |
def extract_payments_csv(): | |
payments_DF = pd.read_csv( | |
'./brazilian-ecommerce/olist_order_payments_dataset.csv') | |
return payments_DF | |
def extract_order_items_csv(): | |
order_items_DF = pd.read_csv( | |
'./brazilian-ecommerce/olist_order_items_dataset.csv') | |
return order_items_DF | |
def extract_customers_csv(): | |
customers_DF = pd.read_csv( | |
'./brazilian-ecommerce/olist_customers_dataset.csv') | |
return customers_DF | |
def extract_orders_csv(): | |
orders_DF = pd.read_csv( | |
'./brazilian-ecommerce/olist_orders_dataset.csv') | |
return orders_DF | |
def extract_sellers_csv(): | |
sellers_DF = pd.read_csv( | |
'./brazilian-ecommerce/olist_sellers_dataset.csv') | |
return sellers_DF | |
def extract_products_csv(): | |
products_DF = pd.read_csv( | |
'./brazilian-ecommerce/olist_products_dataset.csv') | |
return products_DF | |
def extract_product_translations(): | |
products_category_translations_DF = pd.read_csv( | |
'./brazilian-ecommerce/product_category_name_translation.csv', | |
index_col='product_category_name') | |
return products_category_translations_DF | |
def extract_geolocation_csv(): | |
geolocation_DF = pd.read_csv( | |
'./brazilian-ecommerce/olist_geolocation_dataset.csv') | |
return geolocation_DF | |
def extract_marketing_leads_csv(): | |
marketing_leads_DF = pd.read_csv( | |
'./marketing-funnel-olist/olist_marketing_qualified_leads_dataset.csv') | |
return marketing_leads_DF | |
def extract_marketing_closed_deals_csv(): | |
marketing_closed_deals_DF = | |
pd.read_csv('./marketing-funnel-olist/olist_closed_deals_dataset.csv') | |
return marketing_closed_deals_DF | |
def transform_category_names(products_DF, products_category_translations_DF): | |
products_DF['product_category_name'] = products_DF['product_category_name'].apply( | |
lambda x: english_category_name(products_category_translations_DF, x)) | |
return products_DF | |
def transform_orders_dates(orders_DF): | |
orders_DF['order_delivered_customer_date'] = pd.to_datetime( | |
orders_DF['order_delivered_customer_date']) | |
orders_DF['order_purchase_timestamp'] = pd.to_datetime( | |
orders_DF['order_purchase_timestamp']) | |
orders_DF['order_approved_at'] = pd.to_datetime( | |
orders_DF['order_approved_at']) | |
orders_DF['order_delivered_carrier_date'] = pd.to_datetime( | |
orders_DF['order_delivered_carrier_date']) | |
orders_DF['order_estimated_delivery_date'] = pd.to_datetime( | |
orders_DF['order_estimated_delivery_date']) | |
return orders_DF | |
def transform_order_items_payments(orders_DF, payments_DF): | |
return pd.merge(orders_DF, payments_DF, on='order_id', how='inner') | |
def transform_order_items_payments_products(order_list_payments, products_DF): | |
return pd.merge(order_list_payments, products_DF, on='product_id', how='inner') | |
def transform_orders_customers(orders_payments_products, customers_DF): | |
return pd.merge(orders_payments_products, customers_DF, how='inner', on='customer_id') | |
def transform_sellers_marketing_closed(sellers_DF, marketing_closed_deals_DF): | |
return pd.merge(sellers_DF, marketing_closed_deals_DF, how='left', on='seller_id') | |
def transform_sellers(orders_customers_products, sellers_marketing): | |
return pd.merge(orders_customers_products, sellers_marketing, how='inner', on='seller_id') | |
def main(): | |
# extraction | |
payments_DF = client.submit(extract_payments_csv) | |
order_items_DF = client.submit(extract_order_items_csv) | |
sellers_DF = client.submit(extract_sellers_csv) | |
orders_DF = client.submit(extract_orders_csv) | |
products_DF = client.submit(extract_products_csv) | |
customers_DF = client.submit(extract_customers_csv) | |
product_category_translations_DF = client.submit( | |
extract_product_translations) | |
geolocation_DF = client.submit(extract_geolocation_csv) | |
marketing_leads_DF = client.submit(extract_marketing_leads_csv) | |
marketing_closed_deals_DF = client.submit( | |
extract_marketing_closed_deals_csv) | |
# transformation | |
orders_DF_transformed = client.submit( | |
transform_orders_dates, orders_DF) | |
products_DF_transformed = client.submit( | |
transform_category_names, products_DF, product_category_translations_DF) | |
orders_payments = client.submit( | |
transform_order_items_payments, orders_DF_transformed, payments_DF) | |
orders_items_payments = client.submit( | |
transform_order_items_payments, orders_payments, order_items_DF) | |
order_items_payments_products = client.submit( | |
transform_order_items_payments_products, orders_items_payments, products_DF_transformed) | |
orders_customers_products = client.submit( | |
transform_orders_customers, order_items_payments_products, customers_DF) | |
sellers_marketing = client.submit( | |
transform_sellers_marketing_closed, sellers_DF, marketing_closed_deals_DF) | |
sellers_customers_orders_products = client.submit( | |
transform_sellers, sellers_marketing, orders_customers_products) | |
# load | |
olist_data_flat = sellers_customers_orders_products.result( | |
).to_parquet('olist_data_flat', compression='brotli') | |
main() |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
from prefect import task, Flow | |
from dask.distributed import Client, LocalCluster | |
from prefect.engine.executors import DaskExecutor | |
import pandas as pd | |
from english_category_name import * | |
import dataset | |
cluster = LocalCluster() | |
client = Client(cluster) | |
@task | |
def extract_payments_csv(): | |
payments_DF = pd.read_csv('./brazilian-ecommerce/olist_order_payments_dataset.csv') | |
return payments_DF | |
@task | |
def extract_order_items_csv(): | |
order_items_DF = pd.read_csv('./brazilian-ecommerce/olist_order_items_dataset.csv') | |
return order_items_DF | |
@task | |
def extract_customers_csv(): | |
customers_DF = pd.read_csv('./brazilian-ecommerce/olist_customers_dataset.csv') | |
return customers_DF | |
@task | |
def extract_orders_csv(): | |
orders_DF = pd.read_csv('./brazilian-ecommerce/olist_orders_dataset.csv') | |
return orders_DF | |
@task | |
def extract_sellers_csv(): | |
sellers_DF = pd.read_csv('./brazilian-ecommerce/olist_sellers_dataset.csv') | |
return sellers_DF | |
@task | |
def extract_products_csv(): | |
products_DF = pd.read_csv('./brazilian-ecommerce/olist_products_dataset.csv') | |
return products_DF | |
@task | |
def extract_product_translations(): | |
products_category_translations_DF = pd.read_csv('./brazilian-ecommerce/product_category_name_translation.csv', | |
index_col='product_category_name') | |
return products_category_translations_DF | |
@task | |
def extract_geolocation_csv(): | |
geolocation_DF = pd.read_csv('./brazilian-ecommerce/olist_geolocation_dataset.csv') | |
return geolocation_DF | |
@task | |
def extract_marketing_leads_csv(): | |
marketing_leads_DF = pd.read_csv('./marketing-funnel-olist/olist_marketing_qualified_leads_dataset.csv') | |
return marketing_leads_DF | |
@task | |
def extract_marketing_closed_deals_csv(): | |
marketing_closed_deals_DF = pd.read_csv('./marketing-funnel-olist/olist_closed_deals_dataset.csv') | |
return marketing_closed_deals_DF | |
@task | |
def transform_category_names(products_DF, products_category_translations_DF): | |
products_DF['product_category_name'] = products_DF['product_category_name'].apply(lambda x: english_category_name(products_category_translations_DF, x)) | |
return products_DF | |
@task | |
def transform_orders_dates(orders_DF): | |
orders_DF['order_delivered_customer_date'] = pd.to_datetime(orders_DF['order_delivered_customer_date']) | |
orders_DF['order_purchase_timestamp'] = pd.to_datetime(orders_DF['order_purchase_timestamp']) | |
orders_DF['order_approved_at'] = pd.to_datetime(orders_DF['order_approved_at']) | |
orders_DF['order_delivered_carrier_date'] = pd.to_datetime(orders_DF['order_delivered_carrier_date']) | |
orders_DF['order_estimated_delivery_date'] = pd.to_datetime(orders_DF['order_estimated_delivery_date']) | |
return orders_DF | |
@task | |
def connect_sql_file(): | |
db = dataset.connect('sqlite://olist_data.db') | |
return db | |
@task | |
def transform_order_items_payments(orders_DF, payments_DF): | |
return pd.merge(orders_DF, payments_DF, on='order_id', how='inner') | |
@task | |
def transform_order_items_payments_products(order_list_payments, products_DF): | |
return pd.merge(order_list_payments, products_DF, on='product_id', how='inner') | |
@task | |
def transform_orders_customers(orders_payments_products, customers_DF): | |
return pd.merge(orders_payments_products, customers_DF, how='inner', on='customer_id') | |
@task | |
def transform_sellers_marketing_closed(sellers_DF, marketing_closed_deals_DF): | |
return pd.merge(sellers_DF, marketing_closed_deals_DF, how='left', on='seller_id') | |
@task | |
def transform_sellers(orders_customers_products, sellers_marketing): | |
return pd.merge(orders_customers_products, sellers_marketing, how='inner', on='seller_id') | |
def main(): | |
#setup Flow | |
with Flow('olist_ETL') as flow: | |
#extraction | |
payments_DF = extract_payments_csv() | |
order_items_DF = extract_order_items_csv() | |
sellers_DF = extract_sellers_csv() | |
orders_DF = extract_orders_csv() | |
products_DF = extract_products_csv() | |
customers_DF = extract_customers_csv() | |
product_category_translations_DF = extract_product_translations() | |
geolocation_DF = extract_geolocation_csv() | |
marketing_leads_DF = extract_marketing_leads_csv() | |
marketing_closed_deals_DF = extract_marketing_closed_deals_csv() | |
#transformation | |
orders_DF_transformed = transform_orders_dates(orders_DF) | |
products_DF_transformed = transform_category_names(products_DF, product_category_translations_DF) | |
orders_payments = transform_order_items_payments(orders_DF_transformed, payments_DF) | |
orders_items_payments = transform_order_items_payments(orders_payments, order_items_DF) | |
order_items_payments_products = transform_order_items_payments_products(orders_items_payments, products_DF_transformed) | |
orders_customers_products = transform_orders_customers(order_items_payments_products, customers_DF) | |
sellers_marketing = transform_sellers_marketing_closed(sellers_DF, marketing_closed_deals_DF) | |
sellers_customers_orders_products = transform_sellers(sellers_marketing, orders_customers_products) | |
#load | |
olist_data_flat = sellers_customers_orders_products.result.to_parquet('olist_data_flat', compression='brotli') | |
executor = DaskExecutor(address=client.scheduler.address) | |
flow.run(executor=executor) | |
if __name__ == "__main__": | |
main() | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment