綜合運用

Data Engineering 入門

Vincent Vankrunkelsven

Data Engineer, DataCamp

ETL 函式

def extract_table_to_df(tablename, db_engine):
  return pd.read_sql("SELECT * FROM {}".format(tablename), db_engine)


def split_columns_transform(df, column, pat, suffixes):
  # Converts column into str and splits it on pat...


def load_df_into_dwh(film_df, tablename, schema, db_engine):
  return film_df.to_sql(tablename, db_engine, schema=schema, if_exists="replace")


db_engines = { ... } # Needs to be configured
def etl():
  # Extract
  film_df = extract_table_to_df("film", db_engines["store"])
  # Transform
  film_df = split_columns_transform(film_df, "rental_rate", ".", ["_dollar", "_cents"])
  # Load
  load_df_into_dwh(film_df, "film", "store", db_engines["dwh"])

Airflow 複習

Airflow 標誌

工作流程排程器
Python
DAGs

DAG 概念範例

以 task 裝飾器或 operators 定義工作

在 Airflow 用 DAG 排程

from airflow.sdk import dag

@dag(dag_id="sample", start_date=datetime(2024, 1, 1),
     schedule="0 0 * * *")
def sample():
    ...

# .------------------------- minute           (0 - 59)
# | .----------------------- hour             (0 - 23)
# | | .--------------------- day of the month (1 - 31)
# | | | .------------------- month            (1 - 12)
# | | | | .----------------- day of the week  (0 - 6)
# * * * * * <command>
0 * * * * # Every hour at the 0th minute

cf. https://crontab.guru

DAG 定義檔

from airflow.sdk import dag, task

@task(task_id="etl_task")
def etl():
    ...

@dag(dag_id="etl_pipeline",
     start_date=datetime(2024, 1, 1),
     schedule="0 0 * * *")
def etl_pipeline():
    wait_for_table = EmptyOperator(task_id="wait")
    wait_for_table >> etl()

etl_pipeline()

DAG 定義檔

from airflow.sdk import dag, task

...

etl_pipeline()

儲存為 etl_dag.py，路徑 ~/airflow/dags/

Airflow 介面

Airflow 介面螢幕截圖

一起來練習吧！

Data Engineering 入門