From 76f91fdea15070846fdfcd3a5dc6d6171ee7f671 Mon Sep 17 00:00:00 2001
From: Marcin Jakubowski <mjakubowski84@gmail.com>
Date: Mon, 5 Feb 2024 19:07:02 +0100
Subject: [PATCH] Bugfix: support reading decimals encoded as long values

---
 build.sbt                                     |  4 +-
 .../parquet4s/ParquetReaderItSpec.scala       |  8 ---
 .../parquet4s/ParquetReadSupport.scala        |  3 +
 .../parquet4s/DecimalValueSpec.scala          | 60 +++++++++++++++++++
 project/metals.sbt                            |  2 +-
 5 files changed, 66 insertions(+), 11 deletions(-)
 create mode 100644 core/src/test/scala/com/github/mjakubowski84/parquet4s/DecimalValueSpec.scala

diff --git a/build.sbt b/build.sbt
index d92c05e3..7cce79a9 100644
--- a/build.sbt
+++ b/build.sbt
@@ -14,8 +14,8 @@ val akkaLib  = ActorLibCross("-akka", "-akka")
 val pekkoLib = ActorLibCross("-pekko", "-pekko")
 
 ThisBuild / organization := "com.github.mjakubowski84"
-ThisBuild / version := "2.16.0-SNAPSHOT"
-ThisBuild / isSnapshot := true
+ThisBuild / version := "2.15.1"
+ThisBuild / isSnapshot := false
 ThisBuild / scalaVersion := twoThirteen
 
 ThisBuild / javacOptions ++= Seq("-source", "1.8", "-target", "1.8")
diff --git a/core/src/it/scala/com/github/mjakubowski84/parquet4s/ParquetReaderItSpec.scala b/core/src/it/scala/com/github/mjakubowski84/parquet4s/ParquetReaderItSpec.scala
index 0987a4b0..f5ca3b54 100644
--- a/core/src/it/scala/com/github/mjakubowski84/parquet4s/ParquetReaderItSpec.scala
+++ b/core/src/it/scala/com/github/mjakubowski84/parquet4s/ParquetReaderItSpec.scala
@@ -71,14 +71,6 @@ class ParquetReaderItSpec extends AnyFreeSpec with Matchers with TestUtils with
     finally partitioned.close()
   }
 
-  "Attempt to read partitions on non-partitioned data should succeed" in {
-    ParquetWriter.of[I].writeAndClose(Path(tempPath, "file.parquet"), Seq(I(1), I(2)))
-
-    val data = ParquetReader.as[I].read(tempPath)
-    try data.toSeq should be(Seq(I(1), I(2)))
-    finally data.close()
-  }
-
   "Attempt to read inconsistent partitions should fail with an exception" in {
     ParquetWriter.of[I].writeAndClose(Path(tempPath, "a=a1/b=b1/file.parquet"), Seq(I(1)))
     ParquetWriter.of[I].writeAndClose(Path(tempPath, "a=a2/file.parquet"), Seq(I(2)))
diff --git a/core/src/main/scala/com/github/mjakubowski84/parquet4s/ParquetReadSupport.scala b/core/src/main/scala/com/github/mjakubowski84/parquet4s/ParquetReadSupport.scala
index c4dbda5a..cf3dafec 100644
--- a/core/src/main/scala/com/github/mjakubowski84/parquet4s/ParquetReadSupport.scala
+++ b/core/src/main/scala/com/github/mjakubowski84/parquet4s/ParquetReadSupport.scala
@@ -114,6 +114,9 @@ abstract private class ParquetRecordConverter[R <: ParquetRecord[?, R]](schema:
         else value
       record = record.add(name, BinaryValue(rescaled))
     }
+
+    override def addLong(value: Long): Unit =
+      record = record.add(name, BinaryValue(Decimals.binaryFromDecimal(BigDecimal(value, scale, mathContext))))
   }
 
   private class DateTimeConverter(name: String, timeUnit: LogicalTypeAnnotation.TimeUnit)
diff --git a/core/src/test/scala/com/github/mjakubowski84/parquet4s/DecimalValueSpec.scala b/core/src/test/scala/com/github/mjakubowski84/parquet4s/DecimalValueSpec.scala
new file mode 100644
index 00000000..744908d9
--- /dev/null
+++ b/core/src/test/scala/com/github/mjakubowski84/parquet4s/DecimalValueSpec.scala
@@ -0,0 +1,60 @@
+package com.github.mjakubowski84.parquet4s
+
+import org.scalatest.flatspec.AnyFlatSpec
+import org.scalatest.matchers.should.Matchers
+import org.apache.parquet.schema.PrimitiveType
+import org.apache.parquet.schema.LogicalTypeAnnotation
+
+import scala.util.Using
+
+class DecimalValueSpec extends AnyFlatSpec with Matchers {
+
+  case class Data(decimal: BigDecimal)
+  val data = Seq(Data(BigDecimal.decimal(1.2f)))
+
+  "Decimal value" should "read from binary" in {
+
+    val outputFile = InMemoryOutputFile(initBufferSize = 1024)
+
+    // write
+    ParquetWriter
+      .generic(Message(Some("binary-dec"), TypedSchemaDef.decimalSchema("decimal")))
+      .writeAndClose(
+        outputFile,
+        Seq(RowParquetRecord("decimal" -> BinaryValue(Decimals.binaryFromDecimal(BigDecimal.decimal(1.2f)))))
+      )
+
+    val inputFile = outputFile.toInputFile
+
+    // read
+    Using(ParquetReader.as[Data].read(inputFile)) { readData =>
+      readData.toSeq shouldBe data
+    }
+  }
+
+  it should "read from long" in {
+
+    val outputFile = InMemoryOutputFile(initBufferSize = 1024)
+
+    // write
+    ParquetWriter
+      .generic(
+        Message(
+          Some("long-dec"),
+          SchemaDef.primitive(
+            primitiveType         = PrimitiveType.PrimitiveTypeName.INT64,
+            logicalTypeAnnotation = Some(LogicalTypeAnnotation.decimalType(4, 8))
+          )("decimal")
+        )
+      )
+      .writeAndClose(outputFile, Seq(RowParquetRecord("decimal" -> LongValue(12000L))))
+
+    val inputFile = outputFile.toInputFile
+
+    // read
+    Using(ParquetReader.as[Data].read(inputFile)) { readData =>
+      readData.toSeq shouldBe data
+    }
+  }
+
+}
diff --git a/project/metals.sbt b/project/metals.sbt
index cbb25c6a..4c9df445 100644
--- a/project/metals.sbt
+++ b/project/metals.sbt
@@ -2,5 +2,5 @@
 
 // This file enables sbt-bloop to create bloop config files.
 
-addSbtPlugin("ch.epfl.scala" % "sbt-bloop" % "1.5.11")
+addSbtPlugin("ch.epfl.scala" % "sbt-bloop" % "1.5.13")