NoSQL
/ NO-sequel or no-ESS-cue-ell /
NoSQL is the umbrella name for databases that DON'T organize everything into the neat rows-and-columns tables that SQL relies on. Instead of forcing your data into a fixed grid up front, they let each record carry whatever shape it needs — which is handy when your data is messy, changes often, or just doesn't fit a spreadsheet.
It comes in a few flavors. Document stores (like MongoDB) keep each record as a flexible JSON-like blob. Key-value stores (like Redis) are a giant dictionary — hand it a key, get back a value, blazingly fast. There are also wide-column and graph databases for other shapes of problem.
The trade-off is real. By dropping rigid tables you gain flexibility and, often, easier scaling across many machines — but you give up some of the strict consistency and tidy relationships that relational databases hand you for free. The name is a bit misleading: read it as 'not only SQL', not 'no SQL allowed'.
db.users.insertOne({
name: "Mei",
roles: ["admin", "editor"],
address: { city: "Taipei" }
})A document store happily nests lists and objects inside one record — no table required.
NoSQL isn't a replacement for SQL — it's a different tool. Many real systems use both: a relational database for orders, a key-value store for caching.