This guide explains what auto_clean(mode="strict") changes
and what data it may permanently remove.
| Mode | What it applies |
|---|---|
safe |
strip_whitespace only — no data removed |
strict |
strip_whitespace + cast_types + drop_duplicates |
Use ar.profile() and ar.suggest_cleaning() to inspect
what strict mode would change before applying it:
import arnio as ar
frame = ar.read_csv("data.csv")
# Step 1 — preview suggested changes
report = ar.profile(frame)
suggestions = ar.suggest_cleaning(frame)
print(suggestions)
# Step 2 — apply if satisfied
clean = ar.auto_clean(frame, mode="strict")Exact duplicate rows are dropped and cannot be recovered:
# These two rows are identical — one will be dropped
# order_id customer city
# 1002 Prasoon London
# 1002 Prasoon London
clean = ar.auto_clean(frame, mode="strict")Strict mode may cast string columns to int64, float64,
or bool. Leading zeros and mixed-type values can be lost:
# "007" may become 7 after cast
# Mixed-type columns may partially fail
clean = ar.auto_clean(frame, mode="strict")import arnio as ar
frame = ar.read_csv("data.csv")
# Always profile first
report = ar.profile(frame)
print(report.summary())
suggestions = ar.suggest_cleaning(frame)
print(suggestions)
# Apply strict only after reviewing suggestions
clean = ar.auto_clean(frame, mode="strict")README.md— quickstart and auto-clean tutorialARCHITECTURE.md— pipeline and cleaning engine design