How to access a column whose name I cannot access in chained operations

Question

I'll preface this by saying this is a toy example - I do have motivations for doing this, as it sits in the middle of some other chained operations.

I have a DataFrame something like

df
Out[234]: 
            host1  host2  host3
dates                          
2014-02-02      1      3      4
2014-02-03      5      2      1
2014-02-04      2      5      6
2014-02-05      4      6      1
2014-02-06      3      2      1

I am trying to produce a new DataFrame consisting of two columns with the hosts being the index - one column being the values in the last row, the second being whether those values in the last row are greater than 1. My corresponding output should then look like:

newdf
Out[235]: 
dates           2014-02-06  passes
host1                    3    True
host2                    2    True
host3                    1   False

How can I do this with chained operations?

---

Accomplishing the output in and of itself is pretty easy I think, I just did

newdf = df.tail(1).T
newdf['passes'] = newdf.iloc[:, 0] > 1

The reason I'm struggling mightily to do it with chained operations is because as soon as I transpose the tail, the column name becomes of type pandas.tslib.Timestamp,

df.tail(1).T
Out[236]: 
dates  2014-02-06
host1           3
host2           2
host3           1

which I can't seem to access to rename with rename, and so I then can't access it in some boolean operation in assign to create the new "passes" column.

Data:

My toy DataFrame can be generated with

df = pd.DataFrame(dict(dates=pd.date_range('2014-02-02', periods=5),
                       host1=[1, 5, 2, 4, 3], 
                       host2=[3, 2, 5, 6, 2], 
                       host3=[4, 1, 6, 1, 1])).set_index('dates')

Answer 1

You can use lambda expression in assign, where the parameter is the result from previous chained operation:

df.tail(1).T.assign(passes = lambda x: x.iloc[:,0] > 1)

#dates  2014-02-06 00:00:00 passes
#host1                    3   True
#host2                    2   True
#host3                    1  False