Why Indexing Makes Your DB INSANELY Fast #backend #developers

Added: 2026-05-05

[00:00:00]Polkit asked, "A database query takes 2 seconds. After indexing, it takes 50 milliseconds. Same data. So, what was the database doing for 1.95 seconds earlier?"

[00:00:12]And I love this question because it exposes a very common assumption that the same data should take the same amount of effort to fetch. But, that's not how databases work.

[00:00:24]Here is a simple explanation with an analogy.

[00:00:28]Imagine a library with a million books.

[00:00:31]You walk in and ask, "Give me all mystery novels by Agatha Christie."

[00:00:35]If the library has no index, no catalog, no lookup system, nothing, the librarian has to walk through shelves, pick books one by one, check the author, check the genre again and again.

[00:00:50]That's your full table scan.

[00:00:53]But, if the library has an index, now there's a catalog system organized by author, genre. The librarian just looks up Agatha Christie, finds [music] mystery section shelf B12, walks straight there.

[00:01:09]Done.

[00:01:10]Same books, completely different effort.

[00:01:14]So, what actually happened?

[00:01:16]Without an index, your database is doing a full table scan. It reads row by row, page by page, from disk or memory, checking each record against your condition.

[00:01:29]That's your 2 seconds.

[00:01:31]After indexing, you've given the database a data structure, usually a B-tree.

[00:01:37]Now, instead of scanning everything, it does a logarithmic lookup. Jump, narrow, jump, found. That's your 50 milliseconds.