Range-Based Partitioning
What you'll learn: How to divide data across partitions using continuous ranges of key values, and the tradeoffs between efficient range queries and uneven load distribution.
What is Range-Based Partitioning?
Range-based partitioning divides your data by assigning continuous ranges of partition key values to different partitions. Instead of scrambling keys with a hash function, you keep keys in their natural sorted order and split them into segments.
For example, imagine partitioning customer records by last name:
- Partition 1: A–F
- Partition 2: G–M
- Partition 3: N–S
- Partition 4: T–Z
Each partition stores all records whose key falls within its assigned range.
How It Differs from Hash-Based Partitioning
Unlike hash-based partitioning (which scatters keys randomly), range-based partitioning preserves sort order. This is similar to organizing a library by call numbers: books with similar numbers sit together, making it trivial to retrieve all books in a numeric range.
Benefits
- Efficient range queries: Fetching all users born between 1990–1995 only touches one or two partitions
- Natural data locality: Related keys (timestamps, alphabetical names) cluster together
- Scan operations: Sequential reads are fast because data is ordered
The Hotspot Risk
The major drawback: uneven distribution. If your application creates many records with keys in one range (say, usernames starting with "S"), that partition becomes a bottleneck while others sit idle. Real-world data rarely distributes uniformly across alphabetical or numeric ranges.
Key Takeaway: Range-based partitioning enables powerful range queries by preserving key order, but requires careful partition boundary selection to avoid creating hotspots in heavily-accessed ranges.