[SPARK-16975][SQL] Column-partition path starting '_' should be handl…

…ed correctly

Currently, Spark ignores path names starting with underscore `_` and `.`. This causes read-failures for the column-partitioned file data sources whose partition column names starts from '_', e.g. `_col`.

**Before**
```scala
scala> spark.range(10).withColumn("_locality_code", $"id").write.partitionBy("_locality_code").save("/tmp/parquet")
scala> spark.read.parquet("/tmp/parquet")
org.apache.spark.sql.AnalysisException: Unable to infer schema for ParquetFormat at /tmp/parquet20. It must be specified manually;
```

**After**
```scala
scala> spark.range(10).withColumn("_locality_code", $"id").write.partitionBy("_locality_code").save("/tmp/parquet")
scala> spark.read.parquet("/tmp/parquet")
res2: org.apache.spark.sql.DataFrame = [id: bigint, _locality_code: int]
```

Pass the Jenkins with a new test case.

Author: Dongjoon Hyun <[email protected]>

Closes #14585 from dongjoon-hyun/SPARK-16975-PARQUET.

(cherry picked from commit abff92b)
Signed-off-by: Cheng Lian <[email protected]>

sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/PartitioningAwareFileCatalog.scala

@@ -204,6 +204,6 @@ abstract class PartitioningAwareFileCatalog(
 
   private def isDataPath(path: Path): Boolean = {
     val name = path.getName
-    !(name.startsWith("_") || name.startsWith("."))
+    !((name.startsWith("_") && !name.contains("=")) || name.startsWith("."))
   }
 }

sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/fileSourceInterfaces.scala

@@ -363,7 +363,7 @@ private[sql] object HadoopFsRelation extends Logging {
     // We filter everything that starts with _ and ., except _common_metadata and _metadata
     // because Parquet needs to find those metadata files from leaf files returned by this method.
     // We should refactor this logic to not mix metadata files with data files.
-    (pathName.startsWith("_") || pathName.startsWith(".")) &&
+    ((pathName.startsWith("_") && !pathName.contains("=")) || pathName.startsWith(".")) &&
       !pathName.startsWith("_common_metadata") && !pathName.startsWith("_metadata")
   }
 

sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/json/JsonFileFormat.scala

@@ -55,7 +55,7 @@ class JsonFileFormat extends TextBasedFileFormat with DataSourceRegister {
           .getOrElse(sparkSession.sessionState.conf.columnNameOfCorruptRecord)
       val jsonFiles = files.filterNot { status =>
         val name = status.getPath.getName
-        name.startsWith("_") || name.startsWith(".")
+        (name.startsWith("_") && !name.contains("=")) || name.startsWith(".")
       }.toArray
 
       val jsonSchema = InferSchema.infer(

sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetFileFormat.scala

@@ -234,7 +234,8 @@ private[sql] class ParquetFileFormat
     // Lists `FileStatus`es of all leaf nodes (files) under all base directories.
     val leaves = allFiles.filter { f =>
       isSummaryFile(f.getPath) ||
-          !(f.getPath.getName.startsWith("_") || f.getPath.getName.startsWith("."))
+        !((f.getPath.getName.startsWith("_") && !f.getPath.getName.contains("=")) ||
+          f.getPath.getName.startsWith("."))
     }.toArray.sortBy(_.getPath.toString)
 
     FileTypes(

sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala

@@ -17,6 +17,7 @@
 
 package org.apache.spark.sql
 
+import java.io.File
 import java.math.MathContext
 import java.sql.{Date, Timestamp}
 
@@ -2591,6 +2592,14 @@ class SQLQuerySuite extends QueryTest with SharedSQLContext {
       Row(s"$expected") :: Nil)
   }
 
+  test("SPARK-16975: Column-partition path starting '_' should be handled correctly") {
+    withTempDir { dir =>
+      val parquetDir = new File(dir, "parquet").getCanonicalPath
+      spark.range(10).withColumn("_col", $"id").write.partitionBy("_col").save(parquetDir)
+      spark.read.parquet(parquetDir)
+    }
+  }
+
   test("SPARK-16644: Aggregate should not put aggregate expressions to constraints") {
     withTable("tbl") {
       sql("CREATE TABLE tbl(a INT, b INT) USING parquet")