Parse Ambiguous Date Formats with pd.to_datetime in Pandas

Question

Given a noisy date column with mixed formats and corrupt text like `'2026-01-15'`, `'15/02/2026'`, and `'UNKNOWN'`, parse them into datetime and isolate corrupt records.

Accepted Answer

Optimal Solution:
import pandas as pd

# errors='coerce' turns corrupt/unparseable strings into NaT (Not-A-Time)
df['parsed_date'] = pd.to_datetime(df['raw_date'], errors='coerce')

# Isolate corrupt records
corrupt_rows = df[df['parsed_date'].isna()]

Technical Explanation:
`pd.to_datetime(..., errors='coerce')` gracefully converts invalid dates into `pd.NaT` instead of raising exceptions, allowing easy filtering of corrupted rows.

raw_date
2026-01-15
UNKNOWN_DATE

parsed_date
2026-01-15 00:00:00
NULL