Scanamo is a library to make using DynamoDB with Scala simpler and less error-prone.
The main focus is on making it easier to avoid mistakes and typos by leveraging Scala's type system and some higher level abstractions.
libraryDependencies ++= Seq(
"com.gu" %% "scanamo" % "0.9.1"
)Scanamo is published for Scala 2.12 and Scala 2.11
If you've used the Java SDK to access Dynamo, the most familiar way to use Scanamo is via the Scanamo object:
scala> import com.gu.scanamo._
scala> import com.gu.scanamo.syntax._
scala> val client = LocalDynamoDB.client()
scala> import com.amazonaws.services.dynamodbv2.model.ScalarAttributeType._
scala> val farmersTableResult = LocalDynamoDB.createTable(client)("farmer")('name -> S)
scala> case class Farm(animals: List[String])
scala> case class Farmer(name: String, age: Long, farm: Farm)
scala> val putResult = Scanamo.put(client)("farmer")(Farmer("McDonald", 156L, Farm(List("sheep", "cow"))))
scala> Scanamo.get[Farmer](client)("farmer")('name -> "McDonald")
res1: Option[Either[error.DynamoReadError, Farmer]] = Some(Right(Farmer(McDonald,156,Farm(List(sheep, cow)))))The Either represents the possibility that an item might exist, but not be parsable into the given
type, in this case Farmer.
Like all the examples in this README and the Scaladoc, this creates a table, so that it can be checked using sbt-doctest, but the same operations can happily run against pre-existing tables.
Scanamo provides a Table abstraction to reduce noise when defining multiple operations against the same table:
scala> import com.gu.scanamo._
scala> import com.gu.scanamo.syntax._
scala> import cats.syntax.either._
scala> val client = LocalDynamoDB.client()
scala> import com.amazonaws.services.dynamodbv2.model.ScalarAttributeType._
scala> val winnersTableResult = LocalDynamoDB.createTable(client)("winners")('name -> S)
scala> case class LuckyWinner(name: String, shape: String)
scala> def temptWithGum(child: LuckyWinner): LuckyWinner = child match {
| case LuckyWinner("Violet", _) => LuckyWinner("Violet", "blueberry")
| case winner => winner
| }
scala> val luckyWinners = Table[LuckyWinner]("winners")
scala> val operations = for {
| _ <- luckyWinners.putAll(
| Set(LuckyWinner("Violet", "human"), LuckyWinner("Augustus", "human"), LuckyWinner("Charlie", "human")))
| winners <- luckyWinners.scan()
| winnerList = winners.flatMap(_.toOption).toList
| temptedWinners = winnerList.map(temptWithGum)
| _ <- luckyWinners.putAll(temptedWinners.toSet)
| results <- luckyWinners.getAll('name -> Set("Charlie", "Violet"))
| } yield results
scala> Scanamo.exec(client)(operations)
res1: Set[Either[error.DynamoReadError, LuckyWinner]] = Set(Right(LuckyWinner(Charlie,human)), Right(LuckyWinner(Violet,blueberry)))Note that when using Table no operations are actually executed against DynamoDB until exec is called.
It's also possible to make more complex queries:
scala> import com.gu.scanamo._
scala> import com.gu.scanamo.syntax._
scala> val client = LocalDynamoDB.client()
scala> import com.amazonaws.services.dynamodbv2.model.ScalarAttributeType._
scala> val transportTableResult = LocalDynamoDB.createTable(client)("transports")('mode -> S, 'line -> S)
scala> case class Transport(mode: String, line: String)
scala> val transportTable = Table[Transport]("transports")
scala> val operations = for {
| _ <- transportTable.putAll(Set(
| Transport("Underground", "Circle"),
| Transport("Underground", "Metropolitan"),
| Transport("Underground", "Central")
| ))
| tubesStartingWithC <- transportTable.query('mode -> "Underground" and ('line beginsWith "C"))
| } yield tubesStartingWithC.toList
scala> Scanamo.exec(client)(operations)
res1: List[Either[error.DynamoReadError, Transport]] = List(Right(Transport(Underground,Central)), Right(Transport(Underground,Circle)))If you want to update some of the fields of a row, which don't form part of the key,
without replacing it entirely, you can use the update operation:
scala> import com.gu.scanamo._
scala> import com.gu.scanamo.syntax._
scala> val client = LocalDynamoDB.client()
scala> import com.amazonaws.services.dynamodbv2.model.ScalarAttributeType._
scala> val teamTableResult = LocalDynamoDB.createTable(client)("teams")('name -> S)
scala> case class Team(name: String, goals: Int, scorers: List[String], mascot: Option[String])
scala> val teamTable = Table[Team]("teams")
scala> val operations = for {
| _ <- teamTable.put(Team("Watford", 1, List("Blissett"), Some("Harry the Hornet")))
| updated <- teamTable.update('name -> "Watford", set('goals -> 2) and append('scorers -> "Barnes") and remove('mascot))
| } yield updated
scala> Scanamo.exec(client)(operations)
res1: Either[error.DynamoReadError, Team] = Right(Team(Watford,2,List(Blissett, Barnes),None))You can also scan and query indexes with Scanamo. In the following example, there is a
table called transport with a hash key of mode and range key of line and a
global secondary index
called colour-index with only a hash key on the colour attribute:
scala> import com.gu.scanamo._
scala> import com.gu.scanamo.syntax._
scala> case class Transport(mode: String, line: String, colour: String)
scala> val transport = Table[Transport]("transport")
scala> val colourIndex = transport.index("colour-index")
scala> val client = LocalDynamoDB.client()
scala> import com.amazonaws.services.dynamodbv2.model.ScalarAttributeType._
scala> LocalDynamoDB.withTableWithSecondaryIndex(client)("transport", "colour-index")('mode -> S, 'line -> S)('colour -> S) {
| val operations = for {
| _ <- transport.putAll(Set(
| Transport("Underground", "Circle", "Yellow"),
| Transport("Underground", "Metropolitan", "Maroon"),
| Transport("Underground", "Central", "Red")))
| maroonLine <- colourIndex.query('colour -> "Maroon")
| } yield maroonLine.toList
| Scanamo.exec(client)(operations)
| }
res0: List[Either[error.DynamoReadError, Transport]] = List(Right(Transport(Underground,Metropolitan,Maroon)))Scanamo also supports asynchronous calls to Dynamo:
scala> import com.gu.scanamo._
scala> import com.gu.scanamo.syntax._
scala> import scala.concurrent.duration._
scala> import scala.concurrent.ExecutionContext.Implicits.global
scala> val client = LocalDynamoDB.client()
scala> import com.amazonaws.services.dynamodbv2.model.ScalarAttributeType._
scala> val farmersTableResult = LocalDynamoDB.createTable(client)("farm")('name -> S)
scala> case class Farm(animals: List[String])
scala> case class Farmer(name: String, age: Long, farm: Farm)
scala> val farmTable = Table[Farmer]("farm")
scala> val ops = for {
| _ <- farmTable.putAll(Set(
| Farmer("Boggis", 43L, Farm(List("chicken"))),
| Farmer("Bunce", 52L, Farm(List("goose"))),
| Farmer("Bean", 55L, Farm(List("turkey")))
| ))
| bunce <- farmTable.get('name -> "Bunce")
| } yield bunce
scala> concurrent.Await.result(ScanamoAsync.exec(client)(ops), 5.seconds)
res1: Option[Either[error.DynamoReadError, Farmer]] = Some(Right(Farmer(Bunce,52,Farm(List(goose)))))Scanamo uses the DynamoFormat type class to define how to read and write
different types to DynamoDB. Scanamo provides such formats for many common
types, but it's also possible to define a serialisation format for types
which Scanamo doesn't provide. For example to store Joda DateTime objects
as ISO Strings in Dynamo:
scala> import org.joda.time._
scala> import com.gu.scanamo._
scala> import com.gu.scanamo.syntax._
scala> case class Foo(dateTime: DateTime)
scala> val client = LocalDynamoDB.client()
scala> import com.amazonaws.services.dynamodbv2.model.ScalarAttributeType._
scala> val fooTableResult = LocalDynamoDB.createTable(client)("foo")('dateTime -> S)
scala> implicit val jodaStringFormat = DynamoFormat.coercedXmap[DateTime, String, IllegalArgumentException](
| DateTime.parse(_).withZone(DateTimeZone.UTC)
| )(
| _.toString
| )
scala> val fooTable = Table[Foo]("foo")
scala> val operations = for {
| _ <- fooTable.put(Foo(new DateTime(0)))
| results <- fooTable.scan()
| } yield results
scala> Scanamo.exec(client)(operations).toList
res1: List[Either[error.DynamoReadError, Foo]] = List(Right(Foo(1970-01-01T00:00:00.000Z)))Scanamo is licensed under the Apache License, Version 2.0 (the "License"); you may not use this software except in compliance with the License.
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.