How do I find my average row size?

In PostgreSQL run: SELECT pg_relation_size('tablename') / reltuples AS avg_row_bytes FROM pg_class WHERE relname = 'tablename'. In MySQL: SELECT AVG_ROW_LENGTH FROM information_schema.TABLES WHERE TABLE_NAME = 'tablename'.

What is a good target partition size?

A common guideline is 1–10 GB per partition for OLTP workloads. Very small partitions increase metadata overhead; very large ones slow partition pruning and maintenance operations.

Does this tool account for indexes?

No — the calculation covers raw row data only. Add an index overhead factor (typically 20–50% of raw data size) in the overhead percentage field to approximate total partition size including indexes.

What is the overhead percentage field for?

Real partitions consume more space than raw row data due to index pages, free-space-map pages, and TOAST tables. A 25% overhead is a reasonable default for most workloads.

Can I use this for NoSQL databases?

Yes. The math is identical for MongoDB collection shards, Cassandra partitions, or Elasticsearch shards — just set average document size as your row size.

Database Partition Size Calculator

Name: Database Partition Size Calculator
Availability: InStock
Author: Nham Vu

Enter your table row count, average row size, and target partition size to get partition count, size per partition, and storage estimates.

Table Parameters

Total Row Count

Number of rows currently in (or expected in) the table.

Average Row Size

Average bytes per row including all columns.

Target Partition Size

Maximum size each partition should reach.

Index / Overhead Factor 25%

Extra space for indexes, free pages, TOAST, etc.

Recommended Partitions

—

Rows per Partition

—

Raw Data Size

—

Total with Overhead

—

Partition Breakdown

Metric	Value

Partition Fill Visualization

—

Fill in the parameters on the left and click Calculate Partitions.

Summary

Enter your table row count, average row size, and target partition size to get partition count, size per partition, and storage estimates.

How it works

Enter the total number of rows in your database table.
Specify the average row size in bytes (or use the byte unit selector).
Set your target partition size — the maximum size each partition should hold.
The calculator outputs the recommended partition count, actual rows per partition, and total storage.
Review the storage breakdown to plan disk capacity and rebalancing schedules.

Use cases

Plan MySQL, PostgreSQL, or MariaDB table partitioning before a large migration.
Estimate how many shards a Cassandra or MongoDB collection needs.
Size partitions for time-series tables where each partition covers a date range.
Capacity-plan disk storage for a growing transactional table.
Validate that existing partition counts are still appropriate after data growth.
Compare partition strategies by adjusting the target size and observing trade-offs.

Frequently Asked Questions

Last updated: 2026-07-22 · Reviewed by Nham Vu