ERR: read_csv with dtype specified on empty data

[Sereger13]

File has this content:

A,A
3,3

When I do pd.read_csv(filepath) - all looks fine, however if I specify skiprows and dtype parameters it fails with the following error:

pd.read_csv(filepath, skiprows=1, dtype=str)

AttributeError: type object 'str' has no attribute 'items'

---

INSTALLED VERSIONS

commit: None
python: 2.7.10.final.0
python-bits: 64
OS: Linux
OS-release: 2.6.18-238.9.1.el5
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US

pandas: 0.17.1

[jreback]

This is equivalent to this. So this doesn't have any data, so the dtype converter should handle this properly.

In [13]: read_csv(StringIO(data),header=0,skiprows=1)
Out[13]: 
Empty DataFrame
Columns: [3, 3.1]
Index: []

In [14]: data
Out[14]: 'A,A\n3,3'

So an even simpler repro is this.

In [16]: read_csv(StringIO('A,B'),dtype=str)
AttributeError: type object 'str' has no attribute 'items'

care to do a pull-request to fix?

[Sereger13]

Prepared the fix locally but getting this error when trying to send pull-request:
'It seems you do not have permission to push your changes to this repository'.

[jreback]

see contributing docs here

you need to fork, then push to your branch and open a PR.

you are trying to push to master