[BUG] Function cudf.DataFrame.pivot doesn't support lists of columns and index arguments

[raisadz]

Describe the bug
Function cudf.DataFrame.pivot doesn't support lists of columns and index arguments. In addition, the error message is misleading.

Steps/Code to reproduce bug

import cudf

data = {'bar': ['x', 'y', 'z', 'w'], 'col': ['a', 'b', 'a', 'b'], 'foo': [1, 2, 3, 4], 'ix': [1, 1, 2, 2]}
df = cudf.DataFrame(data)
df.pivot(columns=['col'], index=['ix'])

This produces the following error:

ValueError                                Traceback (most recent call last)
Cell In[19], line 1
----> 1 df.pivot(columns=['col'], index=['ix'])

File /opt/conda/lib/python3.10/site-packages/cudf/utils/performance_tracking.py:51, in _performance_tracking.<locals>.wrapper(*args, **kwargs)
     43 if nvtx.enabled():
     44     stack.enter_context(
     45         nvtx.annotate(
     46             message=func.__qualname__,
   (...)
     49         )
     50     )
---> 51 return func(*args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/cudf/core/dataframe.py:7475, in DataFrame.pivot(self, columns, index, values)
   7472 @_performance_tracking
   7473 @copy_docstring(reshape.pivot)
   7474 def pivot(self, *, columns, index=no_default, values=no_default):
-> 7475     return cudf.core.reshape.pivot(
   7476         self, index=index, columns=columns, values=values
   7477     )

File /opt/conda/lib/python3.10/site-packages/cudf/core/reshape.py:1036, in pivot(data, columns, index, values)
   1034     index = df.index
   1035 else:
-> 1036     index = cudf.core.index.Index(df.loc[:, index])
   1037 columns = cudf.Index(df.loc[:, columns])
   1039 # Create a DataFrame composed of columns from both
   1040 # columns and index

File /opt/conda/lib/python3.10/site-packages/cudf/core/index.py:87, in IndexMeta.__call__(cls, data, *args, **kwargs)
     82     raise NotImplementedError(
     83         "tupleize_cols is currently not supported."
     84     )
     86 if cls is Index:
---> 87     return as_index(
     88         arbitrary=data,
     89         *args,
     90         **kwargs,
     91     )
     92 return super().__call__(data, *args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/cudf/utils/performance_tracking.py:51, in _performance_tracking.<locals>.wrapper(*args, **kwargs)
     43 if nvtx.enabled():
     44     stack.enter_context(
     45         nvtx.annotate(
     46             message=func.__qualname__,
   (...)
     49         )
     50     )
---> 51 return func(*args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/cudf/core/index.py:3114, in as_index(arbitrary, nan_as_null, copy, name, dtype)
   3110     return cudf.MultiIndex.from_pandas(
   3111         arbitrary.copy(deep=copy), nan_as_null=nan_as_null
   3112     )
   3113 elif isinstance(arbitrary, cudf.DataFrame) or is_scalar(arbitrary):
-> 3114     raise ValueError("Index data must be 1-dimensional and list-like")
   3115 else:
   3116     return as_index(
   3117         column.as_column(arbitrary, dtype=dtype, nan_as_null=nan_as_null),
   3118         copy=copy,
   3119         name=name,
   3120         dtype=dtype,
   3121     )

ValueError: Index data must be 1-dimensional and list-like

Expected behavior
Expected the same output as produced by pandas:

print(df.to_pandas().pivot(columns=['col'], index=['ix']))
bar    foo   
col   a  b   a  b
ix               
1     x  y   1  2
2     z  w   3  4

Environment overview (please complete the following information)

• Environment location: Kaggle notebook

Environment details

print(cudf.__version__)
24.08.03

Additional context
Add any other context about the problem here.

[mroeschke]

Thanks for the report!

I opened #17373 to fix this, and hopefully it should be included in the 24.12 release