[SPARK-18292][SQL] LogicalPlanToSQLSuite should not use resource dependent path for golden file generation

## What changes were proposed in this pull request?

`LogicalPlanToSQLSuite` uses the following command to update the existing answer files.

```bash
SPARK_GENERATE_GOLDEN_FILES=1 build/sbt "hive/test-only *LogicalPlanToSQLSuite"
```

However, after introducing `getTestResourcePath`, it fails to update the previous golden answer files in the predefined directory. This issue aims to fix that.

## How was this patch tested?

It's a testsuite update. Manual.

Author: Dongjoon Hyun <dongjoon@apache.org>

Closes #15789 from dongjoon-hyun/SPARK-18292.

1 files changed, 9 insertions, 1 deletions

diff --git a/sql/hive/src/test/scala/org/apache/spark/sql/catalyst/LogicalPlanToSQLSuite.scala b/sql/hive/src/test/scala/org/apache/spark/sql/catalyst/LogicalPlanToSQLSuite.scala
index 12d18dc87c..8696337b9d 100644
--- a/sql/hive/src/test/scala/org/apache/spark/sql/catalyst/LogicalPlanToSQLSuite.scala
+++ b/sql/hive/src/test/scala/org/apache/spark/sql/catalyst/LogicalPlanToSQLSuite.scala
@@ -46,7 +46,15 @@ class LogicalPlanToSQLSuite extends SQLBuilderTest with SQLTestUtils {
 
   // Used for generating new query answer files by saving
   private val regenerateGoldenFiles: Boolean = System.getenv("SPARK_GENERATE_GOLDEN_FILES") == "1"
-  private val goldenSQLPath = getTestResourcePath("sqlgen")
+  private val goldenSQLPath = {
+    // If regenerateGoldenFiles is true, we must be running this in SBT and we use hard-coded
+    // relative path. Otherwise, we use classloader's getResource to find the location.
+    if (regenerateGoldenFiles) {
+      java.nio.file.Paths.get("src", "test", "resources", "sqlgen").toFile.getCanonicalPath
+    } else {
+      getTestResourcePath("sqlgen")
+    }
+  }
 
   protected override def beforeAll(): Unit = {
     super.beforeAll()