From 518cf6eaf0b5093bcfab235167929b55acaa310d Mon Sep 17 00:00:00 2001
From: Eric Hills <53243273+ebhills@users.noreply.github.com>
Date: Sat, 25 Jul 2026 15:10:00 -0500
Subject: [PATCH 1/3] Add value counts aggregation to select.group_by

---
 tests/recipes/wrangles/test_select.py | 21 +++++++++++++++++++++
 wrangles/recipe_wrangles/select.py    | 21 +++++++++++++++++++++
 2 files changed, 42 insertions(+)

diff --git a/tests/recipes/wrangles/test_select.py b/tests/recipes/wrangles/test_select.py
index 813d3f679..834f948f5 100644
--- a/tests/recipes/wrangles/test_select.py
+++ b/tests/recipes/wrangles/test_select.py
@@ -1940,6 +1940,27 @@ def test_group_by_to_list_multiple_list(self):
             list(df.values[1]) == ['b', [4], [8]]
         )
 
+    def test_group_by_value_counts_as_json_safe_dictionary(self):
+        df = wrangles.recipe.run(
+            """
+            wrangles:
+              - select.group_by:
+                  counts:
+                    - Selection: Selection Counts
+                    - Review: Review Counts
+                  auto_rename_columns: false
+            """,
+            dataframe=pd.DataFrame({
+                "Selection": ["Primary", "Primary", "None"],
+                "Review": [False, False, True],
+            }),
+        )
+
+        assert df.to_dict(orient="records") == [{
+            "Selection Counts": {"Primary": 2, "None": 1},
+            "Review Counts": {"false": 2, "true": 1},
+        }]
+
 
     def test_group_by_where(self):
         """
diff --git a/wrangles/recipe_wrangles/select.py b/wrangles/recipe_wrangles/select.py
index d170e83ca..7acf2054c 100644
--- a/wrangles/recipe_wrangles/select.py
+++ b/wrangles/recipe_wrangles/select.py
@@ -318,6 +318,13 @@ def group_by(
           - string
           - array
         description: The count of values for these column(s)
+      counts:
+        type:
+          - string
+          - array
+        description: >-
+          Return a dictionary containing the count of each distinct value for
+          these column(s). Keys are converted to JSON-safe strings.
       std:
         type:
           - string
@@ -399,6 +406,20 @@ def percentile_(x):
         # Add option to group as a list
         elif operation == "list":
             operation = list
+        elif operation == "counts":
+            def counts(values):
+                output = {}
+                for value, count in values.value_counts(dropna=False).items():
+                    if _pd.isna(value):
+                        key = "null"
+                    elif isinstance(value, bool):
+                        key = str(value).lower()
+                    else:
+                        key = str(value)
+                    output[key] = int(count)
+                return output
+            counts.__name__ = "counts"
+            operation = counts
 
         if not isinstance(columns, list): columns = [columns]
         for column in columns:

From f85253101ec897852b6e899d4bbe013fa71be5a9 Mon Sep 17 00:00:00 2001
From: Eric Hills <53243273+ebhills@users.noreply.github.com>
Date: Fri, 14 Aug 2026 15:03:02 -0500
Subject: [PATCH 2/3] Add null values to test case

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
---
 tests/recipes/wrangles/test_select.py | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/tests/recipes/wrangles/test_select.py b/tests/recipes/wrangles/test_select.py
index 834f948f5..dcf2f1a29 100644
--- a/tests/recipes/wrangles/test_select.py
+++ b/tests/recipes/wrangles/test_select.py
@@ -1951,14 +1951,14 @@ def test_group_by_value_counts_as_json_safe_dictionary(self):
                   auto_rename_columns: false
             """,
             dataframe=pd.DataFrame({
-                "Selection": ["Primary", "Primary", "None"],
-                "Review": [False, False, True],
+                "Selection": ["Primary", "Primary", "None", None],
+                "Review": [False, False, True, None],
             }),
         )
 
         assert df.to_dict(orient="records") == [{
-            "Selection Counts": {"Primary": 2, "None": 1},
-            "Review Counts": {"false": 2, "true": 1},
+            "Selection Counts": {"Primary": 2, "None": 1, "null": 1},
+            "Review Counts": {"false": 2, "true": 1, "null": 1},
         }]
 
 

From ae8a5c2c48d21e5d1529b9609d24d0a5cbfdb105 Mon Sep 17 00:00:00 2001
From: Eric Hills <53243273+ebhills@users.noreply.github.com>
Date: Fri, 14 Aug 2026 15:04:07 -0500
Subject: [PATCH 3/3] Document null and bool behavior in schema

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
---
 wrangles/recipe_wrangles/select.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/wrangles/recipe_wrangles/select.py b/wrangles/recipe_wrangles/select.py
index 7acf2054c..dc4f18377 100644
--- a/wrangles/recipe_wrangles/select.py
+++ b/wrangles/recipe_wrangles/select.py
@@ -324,7 +324,8 @@ def group_by(
           - array
         description: >-
           Return a dictionary containing the count of each distinct value for
-          these column(s). Keys are converted to JSON-safe strings.
+          these column(s). Keys are converted to JSON-safe strings; missing
+          values use the key "null" and booleans use lowercase "true"/"false".
       std:
         type:
           - string