fix: pandas bug when data is blank on post-processing (apache#20629)

* fix pandas bug when data is blank on post-processing

* account for multiple queries when data is blank

(cherry picked from commit c2be54c)

superset/charts/post_processing.py

@@ -324,12 +324,17 @@ def apply_post_process(
     post_processor = post_processors[viz_type]
 
     for query in result["queries"]:
+        if query["result_format"] not in (rf.value for rf in ChartDataResultFormat):
+            raise Exception(f"Result format {query['result_format']} not supported")
+
+        if not query["data"]:
+            # do not try to process empty data
+            continue
+
         if query["result_format"] == ChartDataResultFormat.JSON:
             df = pd.DataFrame.from_dict(query["data"])
         elif query["result_format"] == ChartDataResultFormat.CSV:
             df = pd.read_csv(StringIO(query["data"]))
-        else:
-            raise Exception(f"Result format {query['result_format']} not supported")
 
         processed_df = post_processor(df, form_data, datasource)