In Cloud Firestore there are three write operations:
add()set()update()In the docs it says that using set(object, { merge: true }) will merge the given object with the existing document.
The same happens when you use update(object)... so what is the difference? It seems strange that google would duplicate functionality like this.
The way I understood the difference:
set without merge will overwrite a document or create it if it doesn't exist yet
set with merge will update fields in the document or create it if it doesn't exists
update will update fields but will fail if the document doesn't exist
create will create the document but fail if the document already exists
There's also a difference in the kind of data you provide to set and update.
For set you always have to provide document-shaped data:
set( {a: {b: {c: true}}}, {merge: true} ) With update you can also use field paths for updating nested values:
update({ 'a.b.c': true }) 
create method in the API?
set operates on document-shaped data, where update takes field path and value pairs. This means you can make changes to deeply nested values with update that are more cumbersome with set. For example: set({a: {b: {c: true}}}, {merge: true}) vs update('a.b.c', true).
Another difference (extending Scarygami's answer) between "set with merge" and "update", is when working with a nested values.
if you have a document structured like this:
{ "friends": { "friend-uid-1": true, "friend-uid-2": true, } } and want to add {"friend-uid-3" : true}
using this:
db.collection('users').doc('random-id').set({ "friends": { "friend-uid-3": true } },{merge:true})
will result in this data:
{ "friends": { "friend-uid-1": true, "friend-uid-2": true, "friend-uid-3": true } } however update using this:
db.collection('users').doc('random-id').update({ "friends": { "friend-uid-3": true } })
will result in this data:
`{ "friends": { "friend-uid-3": true } }` 
{ merge: true } to the update function.
update, you can use dot notation. The overwrite behavior of update is different if you do/don't use dot notation.Per docs: https://firebase.google.com/docs/firestore/manage-data/add-data#update_fields_in_nested_objects
Dot notation allows you to update a single nested field without overwriting other nested field. If you update a nested field without dot notation, you will overwrite the entire map field.
As stated above, this replaces entire friends structure.
db.collection('users').doc('random-id').update({ "friends": { "friend-uid-3": true } }) This does not.
db.collection('users').doc('random-id').update({ "friends.friend-uid-3": true }) Further adding on to the answers above, if you want to delete nested fields in a map then you may want to use update or set depending on your use case.
If you start with the following and want to remove all profile entries other than "user1" then you have two options.
{ "users": { "profiles": { "user1": ..., "user2": ... } } This will overwrite profiles with whatever is provided
update({ 'users.profiles': { 'user1': ... } }) This will merge the deletes into the existing profiles, leaving whatever wasn't deleted
set({ users: { profiles: { 'user2': FieldValue.delete(), 'user3': FieldValue.delete(), ... } } }, { merge: true }) This only applies to Maps because both set and update will overwrite arrays unless you explicitly use the array-specific operators such as arrayUnion.
Scarygami answer holds true for firebase version 9 as well. Small changes should be considered though:
setDoc instead of set. For example, instead of using db.collection('users').doc('random-id').set({ 'admin': true });, use setDoc(doc(db, 'users', 'random-id'), { "admin": true })updateDoc instead of update. For example, instead of using db.collection('users').doc( random-id').update({"age": 27}), use await updateDoc(doc(db, 'users', 'random-id'), { "age": 27})