BUG: DataFrame.agg and apply with 'size' returns a scalar

Compare

```
df = pd.DataFrame({'a': [1, 2, 3], 'b': [4, 5, 6]})
print(df.apply('size'))
print(df.apply(['size']))
print('---')
print(df.agg('size'))
print(df.agg(['size']))
```

gives

```
6
      a  b
size  3  3
---
6
      a  b
size  3  3
```

agg/apply should operate on rows/columns.