fix: pivot columns with ints for name (apache#16259)

(cherry picked from commit 9b2dffe)

superset/charts/post_processing.py

@@ -177,7 +177,7 @@ def pivot_table_v2(  # pylint: disable=too-many-branches
         df = df.stack(0).unstack().reindex(level=-1, columns=metrics)
 
     # flatten column names
-    df.columns = [" ".join(column) for column in df.columns]
+    df.columns = [" ".join(str(name) for name in column) for column in df.columns]
 
     return df
 

superset/reports/commands/execute.py

@@ -182,6 +182,7 @@ def _get_screenshot(self) -> bytes:
         screenshot: Optional[BaseScreenshot] = None
         if self._report_schedule.chart:
             url = self._get_url(standalone="true")
+            logger.info("Screenshotting chart at %s", url)
             screenshot = ChartScreenshot(
                 url,
                 self._report_schedule.chart.digest,
@@ -190,6 +191,7 @@ def _get_screenshot(self) -> bytes:
             )
         else:
             url = self._get_url()
+            logger.info("Screenshotting dashboard at %s", url)
             screenshot = DashboardScreenshot(
                 url,
                 self._report_schedule.dashboard.digest,
@@ -235,6 +237,7 @@ def _get_csv_data(self) -> bytes:
                 ) from ex
 
         try:
+            logger.info("Getting chart from %s", url)
             csv_data = get_chart_csv_data(url, auth_cookies)
         except SoftTimeLimitExceeded:
             raise ReportScheduleCsvTimeout()