Benchmark for Read performance

Author: jatcwang

build.sbt

@@ -496,7 +496,7 @@ lazy val bench = project
   .enablePlugins(NoPublishPlugin)
   .enablePlugins(AutomateHeaderPlugin)
   .enablePlugins(JmhPlugin)
-  .dependsOn(core, postgres)
+  .dependsOn(core, postgres, hikari)
   .settings(doobieSettings)
 
 lazy val docs = project

modules/bench/src/main/scala/doobie/bench/large.scala

@@ -0,0 +1,157 @@
+// Copyright (c) 2013-2020 Rob Norris and Contributors
+// This software is licensed under the MIT License (MIT).
+// For more information see LICENSE or https://opensource.org/licenses/MIT
+
+package doobie.bench
+
+import cats.effect.IO
+import com.zaxxer.hikari.{HikariConfig, HikariDataSource}
+import doobie.*
+import doobie.syntax.all.*
+import org.openjdk.jmh.annotations.*
+import org.openjdk.jmh.infra.Blackhole
+import scala.util.Using
+
+@State(Scope.Benchmark)
+@OperationsPerInvocation(10000) // We process 10k rows so adjust the benchmark output accordingly
+class LargeRow {
+  import cats.effect.unsafe.implicits.global
+
+  private val hikariConfig = {
+    val config = new HikariConfig()
+    config.setDriverClassName("org.postgresql.Driver")
+    config.setJdbcUrl("jdbc:postgresql:world")
+    config.setUsername("postgres")
+    config.setPassword("password")
+    config.setMaximumPoolSize(2)
+    config
+  }
+
+  val pool = new HikariDataSource(hikariConfig)
+
+  val (xa, cleanup) = {
+    (for {
+      connectEC <- ExecutionContexts.fixedThreadPool[IO](hikariConfig.getMaximumPoolSize)
+    } yield Transactor.fromDataSource[IO].apply[HikariDataSource](pool, connectEC)).allocated.unsafeRunSync()
+  }
+
+  @Setup(Level.Trial)
+  def setup(): Unit = {
+    val connio = for {
+      _ <- sql"""DROP TABLE IF EXISTS data""".update.run
+      _ <- sql"""CREATE TABLE data (
+                  col1 DOUBLE PRECISION,
+                  col2 VARCHAR(50),
+                  col3 INTEGER,
+                  col4 VARCHAR(50),
+                  col5 DOUBLE PRECISION,
+                  col6 DOUBLE PRECISION,
+                  col7 VARCHAR(50),
+                  col8 VARCHAR(50)
+              );""".update.run
+      _ <- sql"select setseed(0.5)".query[Unit].unique // deterministic seed
+      _ <- sql"""INSERT INTO data (col1, col2, col3, col4, col5, col6, col7, col8)
+              SELECT random(), random() :: text, (random() * 1000) :: int, random() :: text, random(), random(), random() :: text, random() :: text
+              FROM generate_series(1, 10000)
+          """.update.run
+    } yield ()
+
+    connio.transact(xa).unsafeRunSync()
+  }
+
+  @TearDown(Level.Trial)
+  def teardown(): Unit = {
+    pool.close()
+    cleanup.unsafeRunSync()
+  }
+
+  @Benchmark
+  def tuple(bh: Blackhole): Unit = {
+    bh.consume(sql"""SELECT col1, col2, col3, col4, col5, col6, col7, col8 FROM data"""
+      .query[(Double, String, Int, String, Double, Double, String, String)].to[List].transact(xa).unsafeRunSync())
+  }
+
+  @Benchmark
+  def tupleOpt(bh: Blackhole): Unit = {
+    bh.consume(sql"""SELECT col1, col2, col3, col4, col5, col6, col7, col8 FROM data"""
+      .query[Option[(Double, String, Int, String, Double, Double, String, String)]].to[List].transact(xa).unsafeRunSync())
+  }
+
+  @Benchmark
+  def semiautoDerivedComplex(bh: Blackhole): Unit = {
+    import SemiautoDerivedInstances.*
+    bh.consume(sql"""SELECT col1, col2, col3, col4, col5, col6, col7, col8 FROM data"""
+      .query[Complex].to[List].transact(xa).unsafeRunSync())
+  }
+
+  @Benchmark
+  def semiautoDerivedComplexOpt(bh: Blackhole): Unit = {
+    import SemiautoDerivedInstances.*
+    bh.consume(sql"""SELECT col1, col2, col3, col4, col5, col6, col7, col8 FROM data"""
+      .query[Option[Complex]].to[List].transact(xa).unsafeRunSync())
+  }
+
+  @Benchmark
+  def autoDerivedComplex(bh: Blackhole): Unit = {
+    import doobie.implicits.*
+    bh.consume(sql"""SELECT col1, col2, col3, col4, col5, col6, col7, col8 FROM data"""
+      .query[Complex].to[List].transact(xa).unsafeRunSync())
+  }
+
+  @Benchmark
+  def rawJdbcComplex(bh: Blackhole): Unit = {
+    var l: List[Complex] = null
+    Using.resource(pool.getConnection()) { c =>
+      Using.resource(c.prepareStatement("SELECT col1, col2, col3, col4, col5, col6, col7, col8 FROM data")) { ps =>
+        Using.resource(ps.executeQuery()) { rs =>
+          val m = scala.collection.mutable.ArrayBuffer.empty[Complex]
+          while (rs.next()) {
+            m += Complex(
+              DSIS(
+                DS(
+                  rs.getDouble(1),
+                  rs.getString(2)
+                ),
+                IS(
+                  rs.getInt(3),
+                  rs.getString(4)
+                )
+              ),
+              DDSS(
+                DD(
+                  rs.getDouble(5),
+                  rs.getDouble(6)
+                ),
+                SS(
+                  rs.getString(6),
+                  rs.getString(7)
+                )
+              )
+            )
+          }
+          l = m.toList
+        }
+      }
+
+    }
+    bh.consume(l)
+  }
+}
+
+case class IS(i: Int, s: String)
+case class DS(d: Double, s: String)
+case class DSIS(ds: DS, is: IS)
+case class DD(d0: Double, d1: Double)
+case class SS(s0: String, s1: String)
+case class DDSS(dd: DD, ss: SS)
+case class Complex(dsis: DSIS, ddss: DDSS)
+
+object SemiautoDerivedInstances {
+  implicit val isRead: Read[IS] = Read.derived
+  implicit val dsRead: Read[DS] = Read.derived
+  implicit val dsisRead: Read[DSIS] = Read.derived
+  implicit val ddRead: Read[DD] = Read.derived
+  implicit val ssRead: Read[SS] = Read.derived
+  implicit val ddssRead: Read[DDSS] = Read.derived
+  implicit val cRead: Read[Complex] = Read.derived
+}