dbt_table_diff

This repository is intended for comparing BigQuery models in dbt that have changed during an open PR.

Usage

The repository has been published as a Github Action and PyPi Package, which means it can be leveraged in a variety of ways:

• Directly in Python via run_dbt_table_diff.
• Directly in Terminal via python3 -m dbt_table_diff.
• In a Github Workflow File via Github Actions to automatically add comments on Open PRs.

Quick Start:

pip3 install dbt_table_diff

Example Code Usage:

from dbt_table_diff import run_dbt_table_diff

run_dbt_table_diff(
        project_id="ultimate-bit-359101",
        keyfile_path="secrets/bq_keyfile.json",
        manifest_file="target/manifest.json",
        dev_prefix="dev_",
        prod_prefix="prod_",
        fallback_prefix="fb_",
        custom_checks_path="",
        ignored_schemas=[],
        irregular_schemas=[],
        org_name="org-not-included",
        repo_name="dbt_example",
        pr_id="2",
        auth_token="my_github_pat",
)

Example CLI Usage:

python3 -m dbt_table_diff -t $GH_TOKEN -o org-not-included -r dbt_example -l 2 \
--manifest_file 'target/manifest.json' --project_id 'ultimate-bit-359101' \
--keyfile_path 'secrets/bq_keyfile.json' --dev_prefix 'dev_' --prod_prefix 'prod_' --fallback_prefix 'fb_'

Example Github Action Usage:

• Overview of Github Actions
• Open PR showing how to use dbt_table_diff as a Github Action.

Github Actions Input Arguments:

Step-By-Step Break Down of Process:

• Fetches list of files modified in Pull Request
  • by CURLing github.api.com/repos/{organization}/{repository}/pulls/{pull_request_id}/files
• Filters on relevant_files
  • which are files matching models/*.sql
• Builds manifest.json
  • By running dbt deps; dbt compile
• Parses manifest.json for relevant_models
  • using manifest-attribute original_file_path matching relevant_files
• Runs all SQL files in dbt_table_diff/sql_checks
  • for each of the relevant_models, compare the two dbt targets (dev_prefix vs prod_prefix)
• Saves output to file
  • in a format supported by Github comments
• Posts comment on open PR
  • leveraging dbt_table_diff PyPi package

Docs

python3 -m dbt_table_diff --help

usage: dbt_table_diff [-h] [-o ORG_NAME] [-r REPO_NAME] [-t AUTH_TOKEN] [-l PR_ID] [--manifest_file MANIFEST_FILE] [--project_id PROJECT_ID] [--keyfile_path KEYFILE_PATH] [--ignored_schemas IGNORED_SCHEMAS]
                      [--irregular_schemas IRREGULAR_SCHEMAS] [--dev_prefix DEV_PREFIX] [--prod_prefix PROD_PREFIX] [--fallback_prefix FALLBACK_PREFIX] [--custom_checks_path CUSTOM_CHECKS_PATH]

optional arguments:
  -h, --help            show this help message and exit
  -o ORG_NAME, --org_name ORG_NAME
                        Owner of GitHub repository.
  -r REPO_NAME, --repo_name REPO_NAME
                        Name of the GitHub repository.
  -t AUTH_TOKEN, --auth_token AUTH_TOKEN
                        User's GitHub Personal Access Token.
  -l PR_ID, --pr_id PR_ID
                        The issue # of the Pull Request.
  --manifest_file MANIFEST_FILE
                        The path to dbt's manifest file.
  --project_id PROJECT_ID
                        The BigQuery Project ID to leverage.
  --keyfile_path KEYFILE_PATH
                        The path to the keyfile to use during BQ calls.
  --ignored_schemas IGNORED_SCHEMAS
                        Folders in models/ to always ignore during row/col checks.
  --irregular_schemas IRREGULAR_SCHEMAS
                        Folders in models/ which use 'fallback_prefix' in prod.
  --dev_prefix DEV_PREFIX
                        Prefix used by development datasets in dbt.
  --prod_prefix PROD_PREFIX
                        Prefix used by production datasets in dbt.
  --fallback_prefix FALLBACK_PREFIX
                        Uncommon prefix used by only some production datasets in dbt.
  --custom_checks_path CUSTOM_CHECKS_PATH
                        A local folder containing any custom SQL to run.