Completion for SQL Server Index Clustering

[pimeys]

On SQL Server a user can choose which one of the indices is defining how the data is clustered to the disk. This means an @id, @unique, @@unique, @@id and @@index attributes can now have an argument clustered with two possible values: true or false.

We'd like to auto-complete these:

model A {
  id Int @id(clustered: true)
//           ^ offer clustered
//                      	^ offer true or false
}

model A {
  id   Int
  id2 Int

  @@id([id, id2], clustered: false)
//                ^ offer clustered
//                           ^ offer true or false
}

And so forth for @unique, @@unique and @@index.

Read the full definition from the proposal document.

Rules:

• Provider must be sqlserver
• Only one id/unique/index can be clustered per model
• By default an id is clustered, a unique key or index are nonclustered

[Jolg42]

Note: Implemented without the Only one id/unique/index can be clustered per model rule, validation makes sure it is enforced with Error parsing attribute "@@unique": A model can only hold one clustered index.