etl（等）

ETL

Extract-Transform-Load data toolkit (v2.0.0). Record and manage data pipeline activities across the full ETL lifecycle — ingest, transform, query, filter, aggregate, visualize, export, sample, schema definition, validation, pipeline orchestration, and data profiling. Each command logs timestamped entries to its own log file, giving you a structured record of all data operations.

Commands

Data Storage

All data is stored as plain-text log files in ~/.local/share/etl/:

• Per-command logs — Each command (ingest, transform, query, etc.) writes to its own .log file (e.g., ingest.log, transform.log).
• History log — Every operation is also appended to history.log with a timestamp and command name.
• Export files — Generated in the same directory as export.json, export.csv, or export.txt.

Entries are stored in timestamp|value format, making them easy to grep, parse, or pipe into downstream tools.

Requirements

• Bash 4.0+ (uses set -euo pipefail)
• coreutils — date, wc, du, head, tail, grep, basename, cut
• No external dependencies, API keys, or network access required
• Works fully offline on any POSIX-compatible system

When to Use

1. Logging data pipeline steps — Record each stage of your ETL process (ingest → transform → validate → export) with timestamps, creating a complete audit trail of data movements.
2. Schema management and validation — Use schema to document table structures and validate to log data quality rules and their pass/fail results.
3. Data profiling and exploration — Use profile to record column statistics, null rates, and distribution anomalies; use sample to log sampling parameters for reproducibility.
4. Pipeline orchestration tracking — Use pipeline to record multi-step workflow configurations, execution order, and dependencies between ETL stages.
5. Cross-team data operations review — Run stats for aggregate counts, search to find specific operations by keyword, and export json to share pipeline logs with team members or load into dashboards.

Examples

# Log a data ingestion from S3
etl ingest "s3://data-lake/raw/users_2024.csv — 1.2M rows, CSV format"

# Record a transformation step
etl transform "Normalize email to lowercase, cast created_at to UTC timestamp"

# Log a validation rule
etl validate "NOT NULL check on user_id: 0 violations out of 1,200,000 rows"

# Record schema for a new table
etl schema "users_dim: id INT PK, email VARCHAR(255), created_at TIMESTAMP, country CHAR(2)"

# Define a pipeline
etl pipeline "daily_user_load: ingest(s3) -> dedupe -> validate -> load(postgres)"

# Search for anything related to 'users'
etl search users

# Export all ETL logs to CSV for analysis
etl export csv

# View summary statistics
etl stats

# Check system health
etl status

Tips

• Run any data command without arguments to see recent entries (e.g., etl ingest shows the last 20 ingest entries).
• Use etl recent for a quick overview of all activity across all categories.
• Combine with cron to auto-log pipeline runs: 0 2 * * * etl pipeline "nightly_load completed at $(date)"
• Back up your data by copying ~/.local/share/etl/ to your preferred backup location.

*Powered by BytesAgain | bytesagain.com | hello@bytesagain.com*