Indexes

A secondary index gives you a second access pattern over the same data — a different partition and sort key. Coming from SQL, the trap is treating a GSI like a B-tree index that's free to add. It isn't: a GSI is a replicated copy of your data with its own keys, its own cost, and its own consistency rules.

This section covers picking the right index type, keeping indexes cheap, and the gotchas that bite in production — stale reads and write throttling that traces back to an index, not the table.

0 of 7 readQuiz

GSI vs LSI in DynamoDB

GSI vs LSI in DynamoDB — the partition-key, consistency, capacity and 10 GB differences, and a simple rule for which secondary index to reach for.

Beginner3 min read

DynamoDB Index Projections: KEYS_ONLY, INCLUDE & ALL

DynamoDB index projections — KEYS_ONLY, INCLUDE, and ALL — what each copies into a GSI/LSI, the storage and write-cost tradeoff, and why a missing attribute forces a base-table fetch.

Intermediate5 min read

DynamoDB Sparse Indexes

DynamoDB sparse indexes let a global secondary index hold only the items that carry its key — the cheapest way to query a small subset of a large table.

Intermediate7 min read

Key Overloading in DynamoDB

Key overloading in DynamoDB — store many entity types under generic PK/SK names, encode the type in the key value, and serve every access pattern from one table.