From 34572272f7d6d5c7994f7e70cdc3c0b408e3e705 Mon Sep 17 00:00:00 2001
From: Emaan Hariri <ehariri@berkeley.edu>
Date: Sat, 16 Apr 2022 10:11:56 -0700
Subject: [PATCH 1/2] change mylen function to account for empty lists

---
 pandas/core/frame.py                       |  2 +-
 pandas/tests/frame/methods/test_explode.py | 46 +++++++++++++++++-----
 2 files changed, 37 insertions(+), 11 deletions(-)

diff --git a/pandas/core/frame.py b/pandas/core/frame.py
index ceede5fdb5577..3a60d3cc19287 100644
--- a/pandas/core/frame.py
+++ b/pandas/core/frame.py
@@ -8616,7 +8616,7 @@ def explode(
         if len(columns) == 1:
             result = df[columns[0]].explode()
         else:
-            mylen = lambda x: len(x) if is_list_like(x) else -1
+            mylen = lambda x: len(x) if is_list_like(x) and len(x) else 1
             counts0 = self[columns[0]].apply(mylen)
             for c in columns[1:]:
                 if not all(counts0 == self[c].apply(mylen)):
diff --git a/pandas/tests/frame/methods/test_explode.py b/pandas/tests/frame/methods/test_explode.py
index 8716a181120f6..437b6cb07fe08 100644
--- a/pandas/tests/frame/methods/test_explode.py
+++ b/pandas/tests/frame/methods/test_explode.py
@@ -223,9 +223,15 @@ def test_explode_sets():
 
 
 @pytest.mark.parametrize(
-    "input_subset, expected_dict, expected_index",
+    "input_dict, input_index, input_subset, expected_dict, expected_index",
     [
         (
+            {
+                "A": [[0, 1, 2], np.nan, [], (3, 4), np.nan],
+                "B": 1,
+                "C": [["a", "b", "c"], "foo", [], ["d", "e"], np.nan],
+            },
+            list("abcde"),
             list("AC"),
             {
                 "A": pd.Series(
@@ -239,6 +245,12 @@ def test_explode_sets():
             list("aaabcdde"),
         ),
         (
+            {
+                "A": [[0, 1, 2], np.nan, [], (3, 4), np.nan],
+                "B": 1,
+                "C": [["a", "b", "c"], "foo", [], ["d", "e"], np.nan],
+            },
+            list("abcde"),
             list("A"),
             {
                 "A": pd.Series(
@@ -260,18 +272,32 @@ def test_explode_sets():
             },
             list("aaabcdde"),
         ),
+        (
+            {
+                "A": [[0, 1, 2], [], np.nan, [], (3, 4), np.nan, []],
+                "B": 1,
+                "C": [["a", "b", "c"], 2, "foo", [], ["d", "e"], np.nan, np.nan],
+            },
+            list("abcdefg"),
+            list("AC"),
+            {
+                "A": pd.Series(
+                    [0, 1, 2, np.nan, np.nan, np.nan, 3, 4, np.nan, np.nan],
+                    index=list("aaabcdeefg"),
+                    dtype=object,
+                ),
+                "B": 1,
+                "C": ["a", "b", "c", 2, "foo", np.nan, "d", "e", np.nan, np.nan],
+            },
+            list("aaabcdeefg"),
+        ),
     ],
 )
-def test_multi_columns(input_subset, expected_dict, expected_index):
+def test_multi_columns(
+    input_dict, input_index, input_subset, expected_dict, expected_index
+):
     # GH 39240
-    df = pd.DataFrame(
-        {
-            "A": [[0, 1, 2], np.nan, [], (3, 4), np.nan],
-            "B": 1,
-            "C": [["a", "b", "c"], "foo", [], ["d", "e"], np.nan],
-        },
-        index=list("abcde"),
-    )
+    df = pd.DataFrame(input_dict, input_index)
     result = df.explode(input_subset)
     expected = pd.DataFrame(expected_dict, expected_index)
     tm.assert_frame_equal(result, expected)

From eed608c2e5d8d7a2e277b2b4546b8f1144a881f5 Mon Sep 17 00:00:00 2001
From: Emaan Hariri <ehariri@berkeley.edu>
Date: Thu, 21 Apr 2022 10:46:27 -0700
Subject: [PATCH 2/2] Clean up mylen and add github issue # to test

---
 pandas/core/frame.py                       | 2 +-
 pandas/tests/frame/methods/test_explode.py | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/pandas/core/frame.py b/pandas/core/frame.py
index 3a60d3cc19287..b4a5e93aec958 100644
--- a/pandas/core/frame.py
+++ b/pandas/core/frame.py
@@ -8616,7 +8616,7 @@ def explode(
         if len(columns) == 1:
             result = df[columns[0]].explode()
         else:
-            mylen = lambda x: len(x) if is_list_like(x) and len(x) else 1
+            mylen = lambda x: len(x) if is_list_like(x) and len(x) > 0 else 1
             counts0 = self[columns[0]].apply(mylen)
             for c in columns[1:]:
                 if not all(counts0 == self[c].apply(mylen)):
diff --git a/pandas/tests/frame/methods/test_explode.py b/pandas/tests/frame/methods/test_explode.py
index 437b6cb07fe08..a6cfe3e994aaa 100644
--- a/pandas/tests/frame/methods/test_explode.py
+++ b/pandas/tests/frame/methods/test_explode.py
@@ -296,7 +296,7 @@ def test_explode_sets():
 def test_multi_columns(
     input_dict, input_index, input_subset, expected_dict, expected_index
 ):
-    # GH 39240
+    # GH 39240, 46084
     df = pd.DataFrame(input_dict, input_index)
     result = df.explode(input_subset)
     expected = pd.DataFrame(expected_dict, expected_index)