From 1d1f0fb167ca88d83dab6b1f70311aadc8603ffe Mon Sep 17 00:00:00 2001
From: Niranjan Artal <nartal@nvidia.com>
Date: Tue, 3 Dec 2024 15:33:12 -0800
Subject: [PATCH] qualx related changes

---
 .../tools/qualx/preprocess.py                 | 21 ++++++++++++-------
 1 file changed, 13 insertions(+), 8 deletions(-)

diff --git a/user_tools/src/spark_rapids_tools/tools/qualx/preprocess.py b/user_tools/src/spark_rapids_tools/tools/qualx/preprocess.py
index 320615b7c..2961f9f90 100644
--- a/user_tools/src/spark_rapids_tools/tools/qualx/preprocess.py
+++ b/user_tools/src/spark_rapids_tools/tools/qualx/preprocess.py
@@ -109,6 +109,10 @@
         'sqlOp_DeserializeToObject',
         'sqlOp_Exchange',
         'sqlOp_Execute InsertIntoHadoopFsRelationCommand',
+        'sqlOp_Execute InsertIntoHadoopFsRelationCommand csv',
+        'sqlOp_Execute InsertIntoHadoopFsRelationCommand parquet',
+        'sqlOp_Execute InsertIntoHadoopFsRelationCommand orc',
+        'sqlOp_Execute InsertIntoHadoopFsRelationCommand json',
         'sqlOp_Expand',
         'sqlOp_Filter',
         'sqlOp_Generate',
@@ -125,15 +129,16 @@
         'sqlOp_Project',
         'sqlOp_ReusedSort',
         'sqlOp_RunningWindowFunction',
-        'sqlOp_Scan csv ',
+        'sqlOp_Scan csv',
         'sqlOp_Scan ExistingRDD Delta Table Checkpoint',
         'sqlOp_Scan ExistingRDD Delta Table State',
-        'sqlOp_Scan JDBCRelation',
-        'sqlOp_Scan json ',
+        'sqlOp_Scan ExistingRDD',
+        'sqlOp_Scan jdbc',
+        'sqlOp_Scan json',
         'sqlOp_Scan OneRowRelation',
-        'sqlOp_Scan orc ',
-        'sqlOp_Scan parquet ',
-        'sqlOp_Scan text ',
+        'sqlOp_Scan orc',
+        'sqlOp_Scan parquet',
+        'sqlOp_Scan text',
         'sqlOp_SerializeFromObject',
         'sqlOp_Sort',
         'sqlOp_SortAggregate',
@@ -481,8 +486,8 @@ def combine_tables(table_name: str) -> pd.DataFrame:
         'Scan DeltaCDFRelation',
         'Scan ExistingRDD Delta Table Checkpoint',
         'Scan ExistingRDD Delta Table State',
-        'Scan JDBCRelation',
-        'Scan parquet ',  # trailing space is also in default sql op name
+        'Scan jdbc',
+        'Scan parquet',
         # GPU
         'GpuScan parquet',
     ]