MongoDB will throw an error if string indices are being too large.
For SQLite, you can indeed create a unique index on
VARCHAR(VERY_LONG) fields, and probably not throw an error; but are you really willing to accept the consequence?
Even for non-unique indices, probably not all indices helps querying performance that much?
Now that I looked it up, I don't even understand most of the terms.
- Bitmap index
- Dense index
- Sparse index
- Reverse index
- Primary index
- Secondary index
Even for MongoDB and nedb, there are choices of Unique Indices, whether sparse or not.
Now, I am curious on how to design a strongly typed database, with good querying performance, especially if flat and involves minimal joins?
- About being well-typed, maybe Nitrite is a good one?
- Well-defined schema can potentially help with typing, but how? Definitely not SQLite, though.
- I like to work with Node.js, so TypeScript? What about the popular Golang?
- Does well-defined schema help with performance?
Top comments (0)