What happened?

When using cursor.adbc_ingest() to write Polars DataFrames containing decimal values to PostgreSQL, specific decimal values experience issues. For example, 44.123456 saved in a parquet file as Decimal(10.6) becomes 4412.345500 when written to a numeric(10,6) column.

I also tested directly with adbc_ingest function and it behaves the same.

Stack Trace

No response

How can we reproduce the bug?

I made a reproduction repository for this issue: https://github.com/MateiNenciu/polars-decimal-bug

Environment/Setup

Environment

• ADBC PostgreSQL Driver: 1.8.0
• ADBC Driver Manager: 1.8.0
• Polars: 1.34.0b1
• PostgreSQL: 17
• Python: 3.12