Boolean indexing into a series with assignment

[IvanUkhov]

I am curious if the following is the intended behavior:

x = pd.Series([None] * 10)
y = [False] * 3 + [True] * 5 + [False] * 2
x[y] = range(5)
x

# 0     None
# 1     None
# 2     None
# 3        3
# 4        4
# 5        0
# 6        1
# 7        2
# 8     None
# 9     None

The expectation was that it would work as with x.loc[y] = range(5):

# 0    None
# 1    None
# 2    None
# 3       0
# 4       1
# 5       2
# 6       3
# 7       4
# 8    None
# 9    None

If so, where can one read about the logic behind?

LANG : en_US.UTF-8 LOCALE : en_US.UTF-8

pandas : 0.25.3
numpy : 1.17.4
pytz : 2018.4
dateutil : 2.5.0
pip : 19.2.3
setuptools : 41.2.0
Cython : None
pytest : 5.1.2
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : 3.8.0
html5lib : 1.0.1
pymysql : None
psycopg2 : None
jinja2 : 2.8
IPython : 6.5.0
pandas_datareader: None
xlsxwriter : None
pandas_gbq : None
pyarrow : 0.15.1
bs4 : 4.7.1
numexpr : None
xarray : None
fastparquet : None
tables : None
s3fs : None
odfpy : None
pytables : None
xlrd : None
gcsfs : None
xlwt : None
sqlalchemy : 1.3.4
scipy : 1.3.1
openpyxl : None
lxml.etree : 3.8.0
matplotlib : 3.0.3
bottleneck : None

[fujiaxiang]

I believe this is a bug.

A quick investigation shows this is caused by the use of np.putmask to set the values. This is an issue when the values provided (i.e range(5)) has different shape with the entire series (array of 10 elements), and numpy extend and repeats the values, the we are actually using [0, 1, 2, 3, 4, 0, 1, 2, 3, 4] to replace the original series and hence caused this issue.

Will take a look at the fixes.

[fujiaxiang]

take