A library for easier and type-safe use of sqlite in the apple device(ios/macos) environment

SQLiteService

It is a library for easier and type-safe use of sqlite in the apple device(ios/macos) environment.

Installation

Currently, only SPM is supported.

How to use it

The interface of SQLiteService simply consists of open/close + run + migration. The operation that returns the Result type is synchronous, and when the Result is passed to the completion handler, it operates asynchronously. (Synchronous operations of asynchronous operation + run action are executed after the migration operation is finished internally in SQLiteService.)

The run method should be called with a closure of this type: (DataBase) throws -> T indicating what action to take and what the result type is. (DataBase) follows the Connection & DataBase protocol. Please refer to the protocol for which functions are supported. (Instead of using SQLiteService, you can directly handle SQLiteDataBase objects that conforms the Connection & DataBase protocol.)

open and close database

private func openDatabaseAndCloseExample() {
    
    let openResult: Result<Void, Error> = self.service.open(path: self.dbPath)
    print("db open result: \(openResult)")
    
    let closeResult: Result<Void, Error> = self.service.close()
    print("db close result: \(closeResult)")
    
  
    self.service.open(path: self.dbPath) { result in
        print("db open result: \(result)")
    }
    
    self.service.close { result in
        print("db close result: \(result)")
    }
}

Table

Table must be defined with ColumnType and EntityType.
Use ColumnType to specify the column name and stored data type.
EntityType must conform to RowValueType, which means that the query result can be returned as a data model.


// MARK: - TableColumn

public protocol TableColumn: RawRepresentable, CaseIterable where RawValue == String {
    
    var dataType: ColumnDataType { get }
}

// MARK: - RowValuetype

public protocol RowValueType {
    
    init(_ cursor: CursorIterator) throws
}

public protocol Table {
    
    associatedtype EntityType: RowValueType
    associatedtype ColumnType: TableColumn
    ...
}

Below are examples of UserTable and PetTable that store user and pet information.


struct UserTable: Table {
    
    // Define table column
    enum Columns: String, TableColumn {
        case uid
        case name
        case age
        case email
        case phone
        case introduction = "intro"
    
        // define each column data type
        var dataType: ColumnDataType {
            switch self {
            case .uid: return .text([.primaryKey(autoIncrement: false)])
            case .name: return .text([.notNull])
            case .age: return .integer([])
            case .email: return .text([.unique, .notNull])
            case .phone: return .text([])
            case .introduction: return .text([])
            }
        }
    }
    
    // Define Entity that conforms to RowValueType protocol.
    struct Entiity: RowValueType {
        let uid: String
        let name: String
        let age: Int?
        let email: String
        let phone: String?
        let introduction: String?
        
        init(_ cursor: CursorIterator) throws {
            self.uid = try cursor.next().unwrap()
            self.name = try cursor.next().unwrap()
            self.age = cursor.next()
            self.email = try cursor.next().unwrap()
            self.phone = cursor.next()
            self.introduction = cursor.next()
        }
    }
    
    typealias EntityType = Entiity
    typealias ColumnType = Columns
    
    static var tableName: String { "Users" }
    
    // mapping each column with Entity property
    static func scalar(_ entity: Entiity, for column: Columns) -> ScalarType? {
        switch column {
        case .uid: return entity.uid
        case .name: return entity.name
        case .age: return entity.age
        case .email: return entity.email
        case .phone: return entity.phone
        case .introduction: return entity.introduction
        }
    }
}

struct User { 
    let uid: String
    let name: String
    var age: Int?
    let email: String
    var phone: String?
    var introduction: String?
    
    init(dummy index: Int) {
        self.uid = "uid:\(index)"
        self.name = "name:\(index)"
        self.age = index % 2 == 0 ? nil: index
        self.email = "email:\(index)"
        self.introduction = ["hello", "world", "!"].randomElement()
    }
}

extension UserTable.Entiity {
    
    init(_ user: User) {
        self.uid = user.uid
        self.name = user.name
        self.age = user.age
        self.email = user.email
        self.phone = user.phone
        self.introduction = user.introduction
    }
}

// define pet table
struct PetTable: Table {
    
    enum Columns: String, TableColumn {
        case uid
        case ownerID = "owner_id"
        case name
        
        var dataType: ColumnDataType {
            switch self {
            case .uid: return .text([.primaryKey(autoIncrement: false), .notNull])
            case .ownerID: return .text([.notNull])
            case .name: return .text([.notNull])
            }
        }
    }
    
    struct Entity: RowValueType {
        let uid: String
        let ownerID: String
        let name: String
        
        init(_ cursor: CursorIterator) throws {
            self.uid = try cursor.next().unwrap()
            self.ownerID = try cursor.next().unwrap()
            self.name = try cursor.next().unwrap()
        }
        
        init(uid: String, ownerID: String, name: String) {
            self.uid = uid
            self.ownerID = ownerID
            self.name = name
        }
    }
    
    static var tableName: String { "pets" }
    typealias ColumnType = Columns
    typealias EntityType = Entity
    
    static func scalar(_ entity: Entity, for column: Columns) -> ScalarType? {
        switch column {
        case .uid: return entity.uid
        case .ownerID: return entity.ownerID
        case .name: return entity.name
        }
    }
}

Another requirement for tables is to indicate which property of the entity matches each column using the static func scalar(_ entity: Entity, for column: Columns) -> ScalarType? type method.
(Actual data is stored by matching entity property values according to the order of columns following the CaseIterable protocol.)

data manipulation

Here are some basic data manipulation usages.

func testSaveDataUsaga() {
    
    _ = self.service.open(path: self.dbPath)
    
    let users = (0..<10).map{ User(dummy: $0) }
    let entities = users.map{ UserTable.Entiity($0) }
    
    /**
     The  task executed by the run method can be sequentially executed synchronously/asynchronously according to the request method by the serial queue.
     */
     self.service.run(execute: { try $0.insert(UserTable.self, entities: entities) }) { result in
         print("save users result: \(result)")
     }
}

func testLoadDataUsage() {
    
    let users = (0..<10).map{ User(dummy: $0) }
    let entities = users.map{ UserTable.Entiity($0) }
    
    let table = UserTable.self
    
    _ = self.service.open(path: self.dbPath)
    _ = self.service.run(execute: { try $0.insert(table, entities: entities) })
    
    let query1 = table.selectAll{ $0.age > 5 }
    let result1: Result<[UserTable.Entiity], Error> = self.service.run(execute: { try $0.load(query1) })
    print("load users older than 5 result \(result1)")
    
    
    let query2 = table.selectAll{ $0.age > 5 && $0.introduction == "hello" }
    let result2: Result<[UserTable.Entiity], Error> = self.service.run(execute: { try $0.load(query2) })
    print("load result: \(result2)")
    
    let query3 = table.selectAll{ $0.uid == "uid:1" }
    let result3: Result<UserTable.Entiity?, Error> = self.service.run(execute: { try $0.loadOne(query3) })
    print("load user 1 result: \(result3)")
    
    let query4 = table.selectAll()
    let result4: Result<[UserTable.Entiity], Error> = self.service.run(execute: { try $0.load(query4) })
    print("load all users result: \(result4)")
}

func testUpdateUsage() {
    
    let table = UserTable.self
    
    var oldUser = User(dummy: 1)
    oldUser.introduction = "old_introduction"
    _ = self.service.open(path: self.dbPath)
    _ = self.service.run(execute: { try $0.insert(table, entities: [.init(oldUser)]) })
    
    let updateQuery = table.update { [$0.introduction == "newIntro"]}
        .where{ $0.uid == "uid:1" }
    _ = self.service.run(execute: { try $0.update(table, query: updateQuery) })
    
    let selectQuery = table.selectAll{ $0.uid == "uid:1" }
    self.service.run(execute: { try $0.loadOne(selectQuery) }) { (result: Result<UserTable.Entiity?, Error>) in
        guard case let .success(user) = result, let updatedUser = user else { return }
        print("updated user intro: \(updatedUser.introduction)")
    }
}

func testDeleteUsage() {
    let table = UserTable.self
    
    let users = (0..<10).map{ User(dummy: $0) }
    let entities = users.map{ UserTable.Entiity($0) }
    
    _ = self.service.open(path: self.dbPath)
    _ = self.service.run(execute: { try $0.insert(table, entities: entities) })
    
    let deleteQuery = table.delete().where{ $0.uid == "uid:1" }
    self.service.run(execute: { try $0.delete(table, query: deleteQuery) }) { result in
        print("delete result: \(result)")
    }
    
    self.wait(for: [expect], timeout: 0.01)
}

func testJoinQueryUsage() {

    let users = (0..<10).map{ User(dummy: $0) }.map{ UserTable.Entiity($0) }
    let pets: [PetTable.Entity] = [
        .init(uid: "p0", ownerID: "uid:1", name: "foo"),
        .init(uid: "p1", ownerID: "uid:3", name: "bar")
    ]
    
    _ = self.service.open(path: self.dbPath)
    _ = self.service.run(execute: { try $0.insert(UserTable.self, entities: users) })
    _ = self.service.run(execute: { try $0.insert(PetTable.self, entities: pets) })
    
    let allUserQuery = UserTable.selectAll()
    let allpetsQuery = PetTable.selectAll()
    let joinQuery = allUserQuery.innerJoin(with: allpetsQuery, on: { ($0.uid, $1.ownerID) })
    
    typealias UserAndPetPair = (UserTable.Entiity, PetTable.Entity)
    let mapping: (CursorIterator) throws -> UserAndPetPair? = { cursor in
        return (try UserTable.Entiity(cursor), try PetTable.Entity(cursor))
    }
    let result = self.service.run(execute: { try $0.load(joinQuery, mapping: mapping) })
    let petOwnerAndPets = try? result.get()
    print("pet owner and pet: \(petOwnerAndPets)")
}

For more information on how to use it, see unit tests.

As you can see from the readme, this project has a lot of missing features and a lot of room for improvement. Feedback or contributions to the project are always welcome. ?

GitHub

https://github.com/sudopark/SQLiteService