BUG: DataFrame Apply unexpectedly casts types (uint64)

[bdmcfar]

When using apply on a series, pandas is returning a series of int64 even though the target values were uint64.

import numpy as np
import pandas as pd

uintDF = pd.DataFrame(np.uint64([1,2,3,4,5]),columns=['Numbers'])
indexDF = pd.DataFrame([2,3,2,1,2],columns=['Indices'])

def retrieve(targetRow,targetDF):
    val = targetDF['Numbers'].iloc[targetRow]
    return val

resultDF = indexDF['Indices'].apply(retrieve,args=(uintDF,))

print uintDF.dtypes

Numbers uint64
dtype: object

print resultDF.dtypes

dtype('int64')

I'm aware this specific functionality is possible through iloc(columnVector) but my real retrieve function is more complex.

Thank you.

INSTALLED VERSIONS

commit: None
python: 2.7.5.final.0
python-bits: 64
OS: Linux
OS-release: 3.0.38-0.5-default
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8

pandas: 0.15.2
nose: 1.2.1
Cython: 0.21.1
numpy: 1.7.1
scipy: 0.12.0
statsmodels: 0.6.1
IPython: 2.3.1
sphinx: 1.1.3
patsy: 0.3.0
dateutil: 2.3
pytz: 2012j
bottleneck: 0.6.0
tables: 3.1.1
numexpr: 2.4
matplotlib: 1.2.0
openpyxl: 1.6.2
xlrd: 0.9.0
xlwt: 0.7.4
xlsxwriter: None
lxml: 3.1.0
bs4: 4.2.1
html5lib: None
httplib2: None
apiclient: None
rpy2: 2.3.3
sqlalchemy: 0.7.8
pymysql: None
psycopg2: None

[jreback]

dealing with uint64 is somewhat tricky; its not well-supported. In any event you should do this:
(this doesn't fix the problem but is a more efficient way of accessing that doesn't involve chain indexing).

In [16]: type(uintDF.ix[0,'Numbers'])
Out[16]: numpy.uint64

In [17]: type(uintDF.ix[2,'Numbers'])
Out[17]: numpy.uint64

[rhshadrach]

I now get uint64. I believe this was part of #47475, but it looks like tests for groupby may need to be added.