diff --git a/.klyban.csv b/.klyban.csv
index 3a1ad65..881ef01 100644
--- a/.klyban.csv
+++ b/.klyban.csv
@@ -7,6 +7,5 @@
 5,"TODO","semantic tags","Allow tags starting with # to refer to a Github issue if data file is in a git repository having a remote on Github.","github","","2023-07-29T15:44:37.854535"
 6,"TODO","style metadata","Add a METADATA column to host styling tags.","","","2023-07-29T15:45:42.414295"
 9,"TODO","fix add","'add' should prompt for columns content.","","","2023-07-29T15:47:01.948986"
-10,"TODO","fix delete","'delete' print the kanban not updated","","","2023-07-29T15:47:54.063379"
+10,"DONE","fix delete","'delete' print the kanban not updated","","","2023-07-29T17:46:23.078957"
 11,"TODO","fix quotes","After 'add', empty columns got remaining quotes.","","","2023-07-29T15:48:48.804824"
-12,"TODO","test","","UX,data,git,core,code,meta,shit,fuck","","2023-07-29T17:32:18.081371"
diff --git a/klyban.py b/klyban.py
index d17c438..af4b3f5 100644
--- a/klyban.py
+++ b/klyban.py
@@ -186,9 +186,12 @@ def cli(context, **kwargs):
 def show(context, tid):
     """Show a task card (if ID is passed) or the whole the kanban (else)."""
 
+    # Because commands invoked before may alter the table,
+    # we need to reload the data.
+    df = load_data(context)
+
     if tid is None:
         # Show the kanban tables.
-        df = context.obj['data']
         if df.empty:
             print("No task.")
             return
@@ -227,7 +230,6 @@ def show(context, tid):
 
     else: # tid is not None.
         # Show a task card.
-        df = context.obj['data']
         row = df.loc[tid]
 
         console = richconsole.Console()