**ROWS Frame Syntax:**

```sql

-- Physical row counting

FUNCTION() OVER (

ORDER BY columns

ROWS BETWEEN start_boundary AND end_boundary

)

```

**RANGE Frame Syntax:**

```sql

-- Value-based grouping

FUNCTION() OVER (

ORDER BY columns

RANGE BETWEEN start_boundary AND end_boundary

)

```

**Time-based RANGE (PostgreSQL):**

```sql

-- Time interval ranges

FUNCTION() OVER (

ORDER BY timestamp_column

RANGE BETWEEN INTERVAL 'X units' PRECEDING AND CURRENT ROW

)

```

**Common Patterns:**

```sql

-- Moving average (ROWS)

ROWS BETWEEN 6 PRECEDING AND CURRENT ROW -- Exactly 7 rows

-- Cumulative sum (RANGE)

RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW -- All tied values

-- Same-day analysis (RANGE)

RANGE BETWEEN CURRENT ROW AND CURRENT ROW -- All rows with same ORDER BY value

```

ROWS counts physical database rows in their storage/processing order

RANGE groups rows by ORDER BY column values, including all tied values

ROWS gives predictable, exact counts regardless of data values

RANGE can include more or fewer rows than specified due to tie handling

Both require ORDER BY clause to define the sorting for frame boundaries

RANGE supports time intervals (INTERVAL) in PostgreSQL for temporal analysis

Full support for both ROWS and RANGE. RANGE supports INTERVAL for time-based frames.