dbt-athena2

This is a adapter leveraged from this repo to better serve our custom needs. It supports addtional capabilities as below:

• Run on dbt-core version 1.1.x
• Support boto3 session to take the credential from from aws profile name
• On schema change support for new columns added
• Add s3 bucket for storing data instead of randomly writing on staging dir

Quick started

Within your python environment, proceed below step to initate a first project. There will be some prompts at during inital steps, refer Configuring your profile section below to properly set it up.

pip install dbt-athena2
dbt init my_dbt_project
export DBT_PROFILES_DIR=`pwd`
cd my_dbt_project
dbt debug # to test connection
dbt seed # to dump your seed data
dbt run # to run all models
dbt run --select model_name # to run specific model

#...and more...

Basic usage

Model configuration

Below show an example how we configure how our model be configured.

• There are 4 supported materialized modes: view, table, incremental and esphemeral. Details here.
• incremental_strategy supports insert_overwrite and append. If partition is specified, it only overwrites partions available from source data.
• on_schema_change support fail, ignore and append_new_columns only and for only incremental materialization. Details here.
• There are some usefule macro such as run_started_at can be referred from here to enhance the flexibility on the model.

{{ config(
    materialized="incremental",
    partitioned_by=["report_month"],
    incremental_strategy="insert_overwrite",
    on_schema_change="append_new_columns"
) }}

{% set run_date = run_started_at.astimezone(modules.pytz.timezone("Asia/Saigon")).strftime("%Y-%m-%d") %}

select cast(working_day as timestamp) working_day,
sum(spend_balance) spend_balance,
sum(td_balance) td_balance,
sum(gs_balance) gs_balance,
cast(date_format(date_parse('{{ run_date }}', '%Y-%m-%d') - interval '1' month, '%Y%m') as int) report_month
from {{ source('analytics', 'eod_balance') }}
where cast(working_day as date) >= date_trunc('month', cast(date_parse('{{ run_date }}', '%Y-%m-%d')  as date)-interval'2'month)
and cast(working_day as date) < date_trunc('month', cast(date_parse('{{ run_date }}', '%Y-%m-%d')  as date)-interval'1'month)
group by working_day
order by working_day desc

Seed

Under folder seeds, place csv seed file ( eg. c_ecom_rate.csv) and the yaml config (eg. c_ecom_rate.yml) as below example. Then run dbt seed

version: 2

seeds:
  - name: c_ecom_rate
    config:
      enabled: true
      quote_columns: true
      tags: accounting | report

Further notes

• If the workgroup is specified in the profile.yml without s3_staging_dir, it will extract the default s3 ouput attached with that work_group when Override client-side settings enabled.

• The boto3 session inherit from devlopment environment; once deployed, it should be obtained permission as role such as EC2 profile instance or K8S service account role.

• Athena limit ALTER ADD COLUMNS with data type date, recommend to parse it to timestamp or string during implementing the model. Details here.

• Athena not accept the comment like /*, to ignore these auto generated comment from dbt, place this query-comment: null in dbt_project.yml file.

Configuring your profile

A dbt profile can be configured to run against AWS Athena using the following configuration:

Example profiles.yml entry:

athena:
  target: dev
  outputs:
    dev:
      database: awsdatacatalog
      region_name: ap-southeast-1
      aws_profile_name: dl-dev-process
      s3_staging_dir: s3://athena-output-bucket/data_services/
      s3_data_dir: s3://athena-data-bucket/{schema_name}/{table_name}/
      schema: accounting
      type: athena

Additional information

• threads is supported
• database and catalog can be used interchangeably

Running tests

First, install the adapter and its dependencies using make (see Makefile):

make install_deps

Next, configure the environment variables in dev.env to match your Athena development environment. Finally, run the tests using make:

make run_tests

References

• How to structure a dbt project
• dbt best practices