I'm pretty sure I'm missing something obvious. I come from a traditional relational database background.

So every document has an Id. Is that Id needed to get that document back quickly, or is searching by a text field as efficient?

export abstract class FireBaseDoc {
  id: string = ``;
}

export class NameModel extends FireBaseDoc {
  first: string = ``;
  last: string = ``;
  email: string = ``;
}

I'm thinking in this case, I would create an index on email and search for a user that way since that's a unique field.

CodePudding user response：

Every document has an Id. Is that Id needed to get that document back quickly, or is searching by a text field as efficient?

The document ID is essentially a primary key like id field most SQL database examples. You cannot create multiple document with duplicate ID in same collection. If you know the document ID, it's easy to get a single document as shown below:

await db.collection("[col_id]").doc("[doc_id]").get()

If you want to fetch a document by a field's value, then you would require a query:

await db.collection("[col_id]").where("field", "==", "value").get();

// Equivalent to:
// SELECT * FROM col WHERE field == "value";

If you are using any authentication service like Firebase Authentication, I would recommend using user's auth UID as document ID as its relatively easier to fetch a single document by ID instead of using a query and write security rules.

I would create an index on email and search for a user that way since that's a unique field.

Firestore does not support any UNIQUE INDEX feature. Only way to ensure unique value for a field in document is to first to manually check if the value is used already or use it as the document ID.

Checkout Get to know Cloud Firestore series for a detailed explanation.