df summary view should show dtype associated with column, not just overall counts.

[jreback]

So 1st view should have dtype line.....

In [38]: DataFrame(randn(10,2))
Out[38]: 
          0         1
0  0.005503 -0.640638
1 -0.865491  1.448725
2 -1.259692 -0.973965
3  0.439085  0.390064
4  0.100232 -0.322328
5  0.820644 -1.234507
6  1.085050 -0.189603
7  0.553376  1.806583
8 -1.190961 -1.908767
9 -1.133783  0.958868

In [39]: DataFrame(randn(10,2)).info()
<class 'pandas.core.frame.DataFrame'>
Int64Index: 10 entries, 0 to 9
Data columns (total 2 columns):
0    10  non-null values
1    10  non-null values
dtypes: float64(2)

more like

In [39]: DataFrame(randn(10,2)).info()
<class 'pandas.core.frame.DataFrame'>
Int64Index: 10 entries, 0 to 9
Data columns (total 2 columns):
0    10  non-null values [datetime64[ns]]
1    10  non-null values [object]
dtypes: float64(1), object(1)

not so much to see if there are dtype issues, as to get an idea
of the data when the column names aren't clear

[jreback]

thats what I meant :)

agreed..thought this was in 0.11....oh well

[cpcloud]

+1 on this

[jreback]

@cpcloud want to do this?

[cpcloud]

sure

[dragoljub]

So then I have to ask do we even need to repeat 'values'? No extra square brackets and no 'values' looks better IMO.

In [39]: DataFrame(randn(10,2)).info()
<class 'pandas.core.frame.DataFrame'>
Int64Index: 10 entries, 0 to 9
Data columns (total 2 columns):
0    10  non-null datetime64[ns]
1    10  non-null object
dtypes: float64(1), object(1)

[cpcloud]

@dragoljub +1 on removing values and square brackets. @y-p @jreback?

Sure.

[jreback]

is the example just a typo? first column is datetime64 but shows float in the summary?

[cpcloud]

yep ... i corrected it over at the PR.