pivot_table should fail if duplicate entries exist

[ukch]

This looks like #147, but I don't have permission to re-open that.

Consider the following situation:

In [3]: dt
Out[3]: 
   foo  bar  baz
0  1    2    3  
1  4    5    6  
2  1    2    4  

In [4]: pandas.pivot_table(dt, values="baz", rows=["foo"], cols=["bar"])
Out[4]: 
  bar  2    5  
foo            
1      3.5  NaN
4      NaN  6  

Notice the column for foo == 1 && bar == 2 shows a value of 3.5. Either this should be in the resulting table twice, or the command should fail with an error.

[ukch]

(Note: I am using pandas 0.5.0 with Python 2.6)

[wesm]

Pivot tables aggregate the values within a group, by default using mean but you can specify other functions. If you want to reshape, use the pivot instance method (df.pivot('foo', 'bar', 'baz')), which will throw an exception if there are duplicates.