[GitHub] [spark] MaxGekk commented on a change in pull request #24417: [SPARK-27522][SQL][TEST] Test migration from INT96 to TIMESTAMP_MICROS for timestamps in parquet

[GitBox <gi...@apache.org>]

MaxGekk commented on a change in pull request #24417: [SPARK-27522][SQL][TEST] Test migration from INT96 to TIMESTAMP_MICROS for timestamps in parquet
URL: https://github.com/apache/spark/pull/24417#discussion_r277177696
 
 

 ##########
 File path: sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetQuerySuite.scala
 ##########
 @@ -881,6 +882,48 @@ class ParquetQuerySuite extends QueryTest with ParquetTest with SharedSQLContext
     }
   }
 
+  test("Migration from INT96 to TIMESTAMP_MICROS timestamp type") {
+    def testMigration(fromTsType: String, toTsType: String): Unit = {
+      def checkAppend(write: DataFrameWriter[_] => Unit, readback: => DataFrame): Unit = {
+        def data(start: Int, end: Int): Seq[Row] = (start to end).map { i =>
+          val ts = new java.sql.Timestamp(TimeUnit.SECONDS.toMillis(i))
+          ts.setNanos(123456000)
+          Row(ts)
+        }
+        val schema = new StructType().add("time", TimestampType)
+        val df1 = spark.createDataFrame(sparkContext.parallelize(data(0, 1)), schema)
+        withSQLConf(SQLConf.PARQUET_OUTPUT_TIMESTAMP_TYPE.key -> fromTsType) {
+          write(df1.write)
+        }
+        val df2 = spark.createDataFrame(sparkContext.parallelize(data(2, 10)), schema)
+        withSQLConf(SQLConf.PARQUET_OUTPUT_TIMESTAMP_TYPE.key -> toTsType) {
+          write(df2.write.mode(SaveMode.Append))
+        }
+        Seq("true", "false").foreach { vectorized =>
+          withSQLConf(SQLConf.PARQUET_VECTORIZED_READER_ENABLED.key -> vectorized) {
+            checkAnswer(readback, df1.unionAll(df2))
+          }
+        }
+      }
+
+      Seq(false, true).foreach { mergeSchema =>
+        withTempPath { file =>
+          checkAppend(_.parquet(file.getCanonicalPath),
+            spark.read.option("mergeSchema", mergeSchema).parquet(file.getCanonicalPath))
+        }
+
+        withSQLConf(SQLConf.PARQUET_SCHEMA_MERGING_ENABLED.key -> mergeSchema.toString) {
+          val tableName = "parquet_timestamp_migration"
+          withTable(tableName) {
+            checkAppend(_.saveAsTable(tableName), spark.table(tableName))
+          }
+        }
+      }
+    }
+
+    testMigration(fromTsType = "INT96", toTsType = "TIMESTAMP_MICROS")
+    testMigration(fromTsType = "TIMESTAMP_MICROS", toTsType = "INT96")
 
 Review comment:
   @dongjoon-hyun Thank you for looking at this. Purpose of this PR is to check that migration from current logical type for timestamps - `INT96` to `TIMESTAMP_MICROS` is safe ... but we can check `TIMESTAMP_MILLIS` as well though it will require to change the test because `TIMESTAMP_MILLIS` obviously truncates Catalyst's timestamps. 

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org