[Python] Fix DataFrame Series.value_counts with bins

CHANGES.md

@@ -105,6 +105,7 @@
 * (Python) `Duration` built from float seconds now rounds to the nearest microsecond instead of truncating, which could lose a microsecond ([#40263](https://github.com/apache/beam/issues/40263)).
 * (Python) `ReadFromText` with `escapechar` no longer skips a delimiter that directly follows an escaped delimiter, which merged two records into one ([#40459](https://github.com/apache/beam/issues/40459)).
 * (Python) DataFrame API: `agg` with a list that includes `'count'` or `'size'` now labels those rows `count`/`size` instead of `sum` ([#40504](https://github.com/apache/beam/issues/40504)).
+* (Python) DataFrame API: `Series.value_counts(bins=...)` no longer always fails with a `TypeError` ([#40503](https://github.com/apache/beam/issues/40503)).
 * Fixed X (Java/Python) ([#X](https://github.com/apache/beam/issues/X)).
 
 ## Security Fixes

sdks/python/apache_beam/dataframe/frames.py

@@ -2345,11 +2345,18 @@ def value_counts(
           reason="order-sensitive")
 
     if bins is not None:
+      # pandas can't compute bins for the empty proxy, so make the proxy from
+      # a single element of the same type.
+      proxy = self._expr.proxy()
+      proxy = pd.Series([0], dtype=proxy.dtype, name=proxy.name).value_counts(
+          normalize=normalize, bins=bins, dropna=dropna)[:0]
       return frame_base.DeferredFrame.wrap(
           expressions.ComputedExpression(
               'value_counts',
               lambda s: s.value_counts(
-                  normalize=normalize, bins=bins, dropna=dropna)[self._expr],
+                  normalize=normalize, bins=bins, dropna=dropna),
+              [self._expr],
+              proxy=proxy,
               requires_partition_by=partitionings.Singleton(
                   reason=(
                       "value_counts with bin specified requires collecting "

sdks/python/apache_beam/dataframe/frames_test.py

@@ -796,6 +796,15 @@ def test_value_counts_with_nans(self):
             value_counts(dropna=dropna, normalize=normalize),
             df)
 
+  def test_series_value_counts_with_bins(self):
+    s = pd.Series([1, 2, 2, 3, 5, 8, 8, 9, np.nan])
+
+    self._run_test(lambda s: s.value_counts(bins=3), s, nonparallel=True)
+    self._run_test(
+        lambda s: s.value_counts(bins=3, normalize=True), s, nonparallel=True)
+    self._run_test(
+        lambda s: s.value_counts(bins=[0, 5, 10]), s, nonparallel=True)
+
   def test_value_counts_does_not_support_sort(self):
     df = pd.DataFrame({
         'num_legs': [2, 4, 4, 6, np.nan, np.nan],