feat: ak.from_rdataframe should accept a single string 'columns'.

src/awkward/operations/ak_from_rdataframe.py

@@ -8,10 +8,17 @@ def from_rdataframe(rdf, columns):
     Args:
         rdf (`ROOT.RDataFrame`): ROOT RDataFrame to convert into an
             Awkward Array.
-        columns (str or tuple of str): A column or multiple columns to be
+        columns (str or iterable of str): A column or multiple columns to be
             converted to Awkward Array.
 
-    Converts ROOT Data Frame columns into an Awkward Array.
+    Converts ROOT RDataFrame columns into an Awkward Array.
+
+    If `columns` is a string, the return value represents a single RDataFrame column.
+
+    If `columns` is any other iterable, the return value is a record array, in which
+    each field corresponds to an RDataFrame column. In particular, if the `columns`
+    iterable contains only one string, it is still a record array, which has only
+    one field.
 
     See also #ak.to_rdataframe.
     """
@@ -24,7 +31,26 @@ def from_rdataframe(rdf, columns):
 def _impl(data_frame, columns):
     import awkward._connect.rdataframe.from_rdataframe  # noqa: F401
 
-    return ak._connect.rdataframe.from_rdataframe.from_rdataframe(
+    if isinstance(columns, str):
+        columns = (columns,)
+        project = True
+    else:
+        columns = tuple(columns)
+        project = False
+
+    if not all(isinstance(x, str) for x in columns):
+        raise ak._errors.wrap_error(
+            TypeError(
+                f"'columns' must be a string or an iterable of strings, not {columns!r}"
+            )
+        )
+
+    out = ak._connect.rdataframe.from_rdataframe.from_rdataframe(
         data_frame,
         columns,
     )
+
+    if project:
+        return out[columns[0]]
+    else:
+        return out

tests/test_1473-from-rdataframe.py

@@ -335,3 +335,18 @@ def test_to_from_data_frame_rvec_of_rvec_of_rvec():
     )
 
     assert ak_array_in.to_list() == ak_array_out["x"].to_list()
+
+
+def test_to_from_data_frame_columns_as_string():
+    ak_array_in = ak.Array(
+        [[[[1.1]]], [[[2.2], [3.3], [], [4.4]]], [[[], [5.5, 6.6], []]]]
+    )
+
+    data_frame = ak.to_rdataframe({"x": ak_array_in})
+
+    ak_array_out = ak.from_rdataframe(
+        data_frame,
+        columns="x",
+    )
+
+    assert ak_array_in.to_list() == ak_array_out.to_list()