What’s the performance impact of using indexes vs linear scans in IndexedDB?

Question

Harry · Answer

Using indexes vs linear scans in IndexedDB has
massive performance implications, especially as your dataset grows. Here’s a breakdown:
Indexed Lookup vs Linear Scan
Factor
Indexed Lookup
Linear Scan
Time Complexity
O(log n) (B-tree traversal)
O(n) (must inspect all records)
Performance (10K items)
~1–5ms
100ms–1000ms+ (or much worse)
Memory Usage
Minimal (returns only matches)
High (may load all into memory)
Scalability
Excellent
Degrades rapidly
Browser Optimization
Fully optimized path
Least optimized
Real-World Example
Suppose you have 10,000 users and want to find the one with email =
x@example.com.
1. With Index (email indexed)
await db.users.where('email').equals('x@example.com').first();
✅ Indexed B-tree lookup → Fast (~milliseconds)
2. Without Index
await db.users.toArray().then(users =andgt; users.find(u =andgt; u.email === 'x@example.com'));
❌ Loads all 10,000 users into memory → Slow + memory-heavy
Benchmarks (Approximate)
Dataset Size
Indexed Query
Linear Scan
1,000 items
~1–2ms
~20–50ms
10,000 items
~2–5ms
~100–300ms
100,000 items
~5–10ms
1–3 seconds
(Browser- and hardware-dependent; rough estimates)
Why Indexed Queries Are Faster
IndexedDB uses B-trees internally for indexes.B-trees allow fast traversal to matching keys without scanning the full store.Index-based queries go directly to matching entries or a narrow range.
Linear Scan Pitfalls
.toArray() retrieves all records..filter()/.find() works in memory, not on the IndexedDB level.These operations block the main thread longer and consume more memory.
Best Practices
Always index fields that are:
Frequently filtered (where(), equals(), between())Used for sorting (orderBy())Use compound indexes for multi-field queries.Avoid toArray().filter() unless the dataset is tiny.
Summary
Indexed queries in IndexedDB are 10x–1000x faster than linear scans on large datasets.
Always use .where() or .orderBy() on indexed fields for scalable, memory-efficient apps.

interview

What’s the performance impact of using indexes vs linear scans in IndexedDB?

Harry

Can you answer this question?

1 Answers

Indexed Lookup vs Linear Scan

Real-World Example

1. With Index (`email` indexed)

2. Without Index

Benchmarks (Approximate)

Why Indexed Queries Are Faster

Linear Scan Pitfalls

Best Practices

Summary

Liked By

Factor	Indexed Lookup	Linear Scan
Time Complexity	O(log n) (B-tree traversal)	O(n) (must inspect all records)
Performance (10K items)	~1–5ms	100ms–1000ms+ (or much worse)
Memory Usage	Minimal (returns only matches)	High (may load all into memory)
Scalability	Excellent	Degrades rapidly
Browser Optimization	Fully optimized path	Least optimized

Dataset Size	Indexed Query	Linear Scan
1,000 items	~1–2ms	~20–50ms
10,000 items	~2–5ms	~100–300ms
100,000 items	~5–10ms	1–3 seconds