Pandas groupby fails when elements are lists

import pandas as pd
df = pd.DataFrame ({'a' : [[1,2,3]]})
g = df.groupby ('a')
g.groups

TypeError: unhashable type: 'list'

**Comment From: chris-b1**

Nested lists aren't a first-class type in `pandas` - they can sometimes be used, but not as groupby keys, because, as the error message says, they aren't hashable.

If possible, it's better to use a `MultiIndex` or some other multi-dim structure to store nested data.  If necessary, you could also convert the list to a tuple, which is hashable.

```python
In [16]: df['a'] = df['a'].apply(tuple)

In [17]: df.groupby('a').groups
Out[17]: {(1, 2, 3): Int64Index([0], dtype='int64')}