DuckDB vs PostgreSQL

Both extensions share the same core functionality but have platform-specific features.

Feature Matrix

Feature	DuckDB	PostgreSQL
Core Type
monetary type	Yes	Yes
Integer storage (no float errors)	Yes	Yes
70+ currencies	Yes	Yes
Currency symbols parsing	Yes	Yes
Banker’s rounding	Yes	Yes
Operators
Arithmetic (+, -, *, /)	Yes	Yes
Comparison (=, <, >, etc.)	Yes	Yes
Currency mismatch errors	Yes	Yes
Aggregates
SUM	Yes	Yes
AVG	Yes	Yes
MIN/MAX	Yes	Yes
Country/Currency
currency_for_country()	Yes	Yes
countries_for_currency()	Yes	Yes
Exchange Rates
ECB rates	Yes	No
10 exchange rate sources	Yes	No
Arrow operator (->>)	Yes	No
Currency macros (USD, EUR)	Yes	No
Historical conversions	Yes	No
Installation
Community extension	Yes	Build from source
WASM support	Yes	No

When to Use DuckDB

• Analytics workloads: Columnar storage, vectorized execution
• WASM/browser: Interactive documentation, client-side analytics
• Exchange rates: Built-in functions for 10 sources
• Currency conversion: Arrow operator and macros for easy conversions
• Data pipelines: Read/write Parquet, CSV, JSON directly

When to Use PostgreSQL

• Production OLTP: Mature, battle-tested for transactional workloads
• Existing PostgreSQL: Add monetary support to existing databases
• Full SQL compliance: Complete SQL standard support
• Ecosystem: Works with all PostgreSQL tools (pgAdmin, psql, etc.)
• Extensions: Combine with other PostgreSQL extensions

Code Compatibility

Most code is portable between platforms:

Works on Both

-- Type creation
SELECT '100 EUR'::monetary;

-- Arithmetic
SELECT '100 EUR'::monetary + '50 EUR'::monetary;
SELECT '100 EUR'::monetary * 2;

-- Comparisons
SELECT '100 EUR'::monetary > '50 EUR'::monetary;

-- Accessors
SELECT monetary_amount('99.99 EUR'::monetary);
SELECT monetary_currency('100 USD'::monetary);

-- Country/currency lookup
SELECT currency_for_country('US');
SELECT countries_for_currency('EUR');

-- Aggregates
SELECT SUM(amount) FROM orders WHERE monetary_currency(amount) = 'EUR';

DuckDB Only

-- Exchange rate functions
SELECT * FROM ecb_exchange_rates();
SELECT * FROM twelvedata_exchange_rates();

-- Currency conversion
SET monetary_exchange_table = 'exchange_rates';
SELECT '100 EUR'::monetary ->> 'USD';
SELECT USD('100 EUR'::monetary);

-- Historical conversion
SELECT USD_AT('100 EUR'::monetary, '2025-01-15');

Migration Path

PostgreSQL to DuckDB

-- Export from PostgreSQL
COPY (SELECT * FROM transactions) TO '/tmp/transactions.csv' CSV HEADER;

-- Import to DuckDB
CREATE TABLE transactions AS
SELECT * FROM read_csv('/tmp/transactions.csv');

DuckDB to PostgreSQL

-- Export from DuckDB
COPY transactions TO '/tmp/transactions.csv' (HEADER);

-- Import to PostgreSQL
\copy transactions FROM '/tmp/transactions.csv' CSV HEADER

Performance Characteristics

Aspect	DuckDB	PostgreSQL
Bulk analytics	Excellent (columnar)	Good
Single-row OLTP	Moderate	Excellent
Concurrent writes	Limited	Excellent
Memory usage	Higher (columnar)	Lower
Startup time	Faster (embedded)	Slower (server)

Shared Code Architecture

Both extensions share C++ code in the shared/ directory:

• monetary.hpp/cpp - Core Monetary struct
• currency.hpp/cpp - Currency enum and metadata
• country.hpp/cpp - Country-to-currency mapping
• parser.hpp/cpp - String parsing logic

This ensures consistent behavior across both platforms.