Automation and output

Use the same client in CI, shell scripts, and reproducible data checks.

Run statements non-interactively

bash

kaveon --local --data-dir ./warehouse -e "SELECT count(*) FROM orders"
kaveon --server https://engine.example.com -f checks.sql
cat checks.sql | kaveon --server https://engine.example.com --format JSONL

Deterministic formats

text

--format table     aligned terminal output
--format vertical  one row per block
--format csv       comma-separated values
--format tsv       tab-separated values
--format json      one JSON document
--format jsonl     one JSON object per row
--no-header        omit column names
--row-limit 5000   bound returned rows

--ignore-errors continues a batch while preserving a failing exit status. Use --pager never in CI and set NO_COLOR=1 for stable logs.

Inspect performance

EXPLAIN ANALYZE reports the plan, stages, task counters, memory, exchange bytes, rows scanned, spill, and the lane that answered the statement. Keep query ids with your test artifacts.