Code Sample, a copy-pastable example if possible
import pandas as pd
df = pd.DataFrame(data=[1,2,3,3,3,3,4,4,4,5,5,5], columns=['abc'], ).sort_values('abc')
df['Rank'] = df['abc'].rank(method='dense')
df['Rank_Pct']= df['abc'].rank(pct=True, method='dense', )
df['Rank_Pct_Manual']= df['Rank'] / df['Rank'].max()
df.head()
Output:
abc Rank Rank_Pct Rank_Pct_Manual
0 1 1.0 0.083333 0.2
1 2 2.0 0.166667 0.4
2 3 3.0 0.250000 0.6
3 3 3.0 0.250000 0.6
4 3 3.0 0.250000 0.6
5 3 3.0 0.250000 0.6
6 4 4.0 0.333333 0.8
7 4 4.0 0.333333 0.8
8 4 4.0 0.333333 0.8
9 5 5.0 0.416667 1.0
10 5 5.0 0.416667 1.0
11 5 5.0 0.416667 1.0
Problem description
If you chose both the pct=True and method='dense' options of Series.rank, you don't get the expected maximum percentile of 1 if there are repeated values in the Series. This is because the function (e.g., rank_1d_float64()) always divides by the total number of elements in the Series. But in the case of the dense method, we should divide by the maximum rank value.
I'm working on a PR now.
Expected Output
I would expect the values of Rank_Pct and Rank_Pct_Manual to be the same, and that the maximum of both should be 1.
abc Rank Rank_Pct Rank_Pct_Manual
0 1 1.0 0.2 0.2
1 2 2.0 0.4 0.4
2 3 3.0 0.6 0.6
3 3 3.0 0.6 0.6
4 3 3.0 0.6 0.6
5 3 3.0 0.6 0.6
6 4 4.0 0.8 0.8
7 4 4.0 0.8 0.8
8 4 4.0 0.8 0.8
9 5 5.0 1.0 1.0
10 5 5.0 1.0 1.0
11 5 5.0 1.0 1.0
Output of pd.show_versions()
commit: None
python: 3.6.3.final.0
python-bits: 64
OS: Darwin
OS-release: 16.7.0
machine: x86_64
processor: i386
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8
pandas: 0.21.0
pytest: 3.2.3
pip: 9.0.1
setuptools: 36.3.0
Cython: None
numpy: 1.13.3
scipy: 0.19.1
pyarrow: 0.7.1
xarray: None
IPython: 6.2.1
sphinx: None
patsy: None
dateutil: 2.6.1
pytz: 2017.3
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 2.0.2
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: 0.999999999
sqlalchemy: 1.1.15
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None
Comment From: jreback
duplicate of #15630.