From fdfa5d0ad414f1628ceacd158fa22140d7849e51 Mon Sep 17 00:00:00 2001
From: Kevin Hu <kevinhu.sh@gmail.com>
Date: Wed, 24 Jul 2024 14:10:45 +0800
Subject: [PATCH] fix graph bug about second retrieval (#1688)

### What problem does this PR solve?

#1651

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
---
 graph/component/base.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/graph/component/base.py b/graph/component/base.py
index b424d226f..06574e32d 100644
--- a/graph/component/base.py
+++ b/graph/component/base.py
@@ -470,7 +470,10 @@ class ComponentBase(ABC):
             break
 
         if upstream_outs:
-            return pd.concat(upstream_outs, ignore_index=False)
+            df = pd.concat(upstream_outs, ignore_index=True)
+            if "content" in df:
+                df = df.drop_duplicates(subset=['content']).reset_index(drop=True)
+            return df
         return pd.DataFrame()
 
     def get_stream_input(self):