Pure PL/pgSQL implementation of iCalendar RRULE (RFC 5545) for PostgreSQL. No C extensions, no compilation, works everywhere.

This library provides complete RRULE recurrence calculation functionality using pure PL/pgSQL. Use this implementation for all environments (local, development, staging, production) to ensure consistency.

Key Features:

• ✅ No C extensions required - Pure PL/pgSQL, works on any PostgreSQL
• ✅ Full timezone support with DST handling - Wall-clock time preserved ("10 AM stays 10 AM" across DST transitions)
• ✅ Consistent everywhere - Same implementation across all environments
• ✅ Production-ready - Comprehensive test suite with 187 tests (including table operation integration tests)
• ✅ RFC 5545 & RFC 7529 compliant - Supports standard RRULE patterns plus SKIP/RSCALE
• ✅ 50-75x faster than Node.js - Excellent performance without compilation
• ✅ Works on managed services - AlloyDB, RDS, Azure Database for PostgreSQL

Computation where your data lives delivers performance impossible with external processing:

• Set-based operations: JOIN recurrence rules against events, bookings, or any table without data transfer
• Native WHERE/aggregation: Filter by occurrence dates, COUNT occurrences, GROUP BY - all in SQL
• Batch processing: Process 100+ schedules in a single query without round trips
• Memory-efficient streaming: SETOF returns results incrementally, not all at once
• 50-75x faster than Node.js for single-schedule operations, infinitely faster for multi-schedule batch queries

See Example Usage for practical patterns including subscription billing, batch updates, and conflict detection.

Pure PL/pgSQL means:

• ✅ No C compiler or build tools required
• ✅ Install with a single SQL file - no configuration
• ✅ Consistent behavior across all environments (dev, staging, production)

Universal compatibility:

• ✅ Google AlloyDB - High-performance PostgreSQL with AI integration
• ✅ Google Cloud SQL - Fully managed PostgreSQL service
• ✅ Amazon Aurora PostgreSQL - Serverless PostgreSQL with automatic scaling
• ✅ Amazon RDS PostgreSQL - Managed PostgreSQL without custom extensions
• ✅ Azure Database for PostgreSQL - Fully supported managed service
• ✅ Self-hosted PostgreSQL - No special configuration required
• ✅ Docker PostgreSQL - Standard postgres image compatibility

npm install rrule-plpgsql

Then install in your database using your ORM/client. See Installation Guide for TypeScript/ORM integration.

# Install via psql
psql -d your_database -f src/install.sql

# Or use curl for one-line install
curl -sL https://raw.githubusercontent.com/sirrodgepodge/rrule_plpgsql/main/src/install.sql | psql -d your_database

Next Steps: See Example Usage for practical patterns.

✅ Production-Ready (Always Enabled):

• FREQ=DAILY - Daily recurrence with date/time filters
• FREQ=WEEKLY - Weekly recurrence with day-of-week selection
• FREQ=MONTHLY - Monthly recurrence with day/week-of-month selection
• FREQ=YEARLY - Yearly recurrence with month/day/week selection

⚠️ Advanced (Optional, Disabled by Default):

• FREQ=HOURLY - Hourly recurrence (8,760/year max)
• FREQ=MINUTELY - Minutely recurrence (525,600/year max)
• FREQ=SECONDLY - Secondly recurrence (31M/year max)

See Sub-Day Operations Guide for enabling these safely.

• COUNT - Limit number of occurrences
• UNTIL - End date for recurrence
• INTERVAL - Frequency multiplier (every N days/weeks/months)
• BYDAY - Filter by day of week (MO, TU, WE, TH, FR, SA, SU)
• BYDAY with ordinals - Positioned weekdays (2MO = 2nd Monday, -1FR = last Friday)
• BYMONTHDAY - Filter by day of month (1-31, -1 = last day)
• BYMONTH - Filter by month (1-12)
• BYYEARDAY - Filter by day of year (1-366, negative supported)
• BYWEEKNO - Filter by ISO week number (1-53, YEARLY only)
• BYSETPOS - Select specific positions from generated set
• WKST - Week start day (SU, MO, TU, WE, TH, FR, SA)
• SKIP - Invalid date handling (OMIT, BACKWARD, FORWARD)
• TZID - Timezone specification with automatic DST handling
• RSCALE - Calendar system (GREGORIAN supported)

See RFC Spec Compliance for complete feature support matrix and examples.

Full RFC 5545 timezone support with automatic DST handling:

-- Meeting stays at 10 AM wall-clock time across DST boundary
SELECT * FROM rrule.all(
    'FREQ=DAILY;COUNT=3;TZID=America/New_York',
    '2025-03-08 10:00:00'::TIMESTAMP
);
-- Returns: March 8 (EST), March 9 (EDT), March 10 (EDT)
-- All at 10:00 AM wall-clock time

See API Reference for TIMESTAMPTZ API details.

psql -d your_database -f src/install.sql

Includes:

• DAILY, WEEKLY, MONTHLY, YEARLY frequencies
• All standard modifiers and filters
• Timezone support with DST handling
• Secure by default (no DoS vectors)

psql -d your_database -f src/install_with_subday.sql

Includes standard PLUS:

• HOURLY, MINUTELY, SECONDLY frequencies
• ⚠️ Security warnings during installation
• Requires application-level validation

See Sub-Day Operations Guide before using this installation.

Error: "Invalid RRULE: FREQ parameter is required"

• Every RRULE must start with FREQ
• Fix: Add FREQ=DAILY (or WEEKLY/MONTHLY/YEARLY)

Error: "COUNT and UNTIL are mutually exclusive"

• Cannot use both COUNT and UNTIL in same RRULE
• Fix: Use either COUNT=10 OR UNTIL=20251231T235959

Error: "BYWEEKNO can only be used with FREQ=YEARLY"

• BYWEEKNO requires YEARLY frequency
• Fix: Change to FREQ=YEARLY;BYWEEKNO=10

See Validation Rules for complete error reference.

All functions support both TIMESTAMP and TIMESTAMPTZ with automatic timezone handling:

-- Generate occurrences
rrule.all(rrule, dtstart) → SETOF TIMESTAMP
rrule.between(rrule, dtstart, start, end) → SETOF TIMESTAMP

-- Query occurrences
rrule.after(rrule, dtstart, after_date) → TIMESTAMP
rrule.before(rrule, dtstart, before_date) → TIMESTAMP
rrule.next(rrule, dtstart) → TIMESTAMP
rrule.most_recent(rrule, dtstart) → TIMESTAMP

-- Utilities
rrule.count(rrule, dtstart) → INTEGER
rrule.overlaps(dtstart, dtend, rrule, mindate, maxdate) → BOOLEAN

-- Every Monday for 4 weeks
SELECT * FROM rrule.all(
    'FREQ=WEEKLY;BYDAY=MO;COUNT=4',
    '2025-01-06 10:00:00'::TIMESTAMP
);

-- With timezone support
SELECT * FROM rrule.all(
    'FREQ=DAILY;COUNT=5;TZID=America/New_York',
    '2025-03-08 10:00:00'::TIMESTAMP  -- DST handled automatically
);

See API Reference for complete function signatures, timezone handling details, and advanced examples.

Contributions welcome! Please:

1. Run all tests (all 10 test suites must pass)
2. Add test coverage for new features
3. Update documentation
4. Follow RFC 5545/7529 specifications
5. Submit pull request

See Development Guide for contribution guidelines and test suite details.

Single schedule: 50-75x faster than Node.js rrule.js Multi-schedule queries: Eliminates application-database round trips entirely

Scalability advantages:

• Set-based batch operations: Query 100+ schedules in parallel without loops
• Streaming results: SETOF returns incrementally, not all at once (constant memory)
• Query planner optimization: PostgreSQL optimizes JOINs and filters with occurrence expansion
• Early-exit optimizations: Stops computation when COUNT/UNTIL limits reached

See Performance Guide for optimization strategies and Development Guide for benchmarking details.

• PostgreSQL 12 or higher
• No C extensions required
• No external dependencies
• Works on all PostgreSQL-compatible platforms

See Development Guide for development setup instructions.