Indices of df.describe(percentiles=np.linspace(...)) are mix of integers and decimal numbers

[jkovacevic]

Code Sample, a copy-pastable example if possible

import numpy as np
import pandas as pd
df = pd.DataFrame(columns=["a"], data=np.random.randint(10, size=30).reshape(-1, 1))
pct = np.linspace(0, 1, 20 + 1)
desc = df.describe(percentiles=pct)

print(desc.index)
print(desc.loc['25%'])
print(desc.loc['30%'])

Problem description

If desc.index is printed out, some percentile indexes are written as decimal numbers (e.g. 30.0%, 55.0%, 85.0%), while others are whole numbers (e.g. 35%, 70%, 100%). Printing percentiles of distribution therefore requires hardcoding some decimal numbers, e.g.

print(desc.loc['0%'])
print(desc.loc['5%'])
print(desc.loc['10%'])
print(desc.loc['15.0%'])
print(desc.loc['20%'])
print(desc.loc['25%'])
print(desc.loc['30.0%'])
print(desc.loc['35%'])
print(desc.loc['40%'])
print(desc.loc['45%'])
print(desc.loc['50%'])
print(desc.loc['55.0%'])
print(desc.loc['60.0%'])
print(desc.loc['65%'])
print(desc.loc['70%'])
print(desc.loc['75%'])
print(desc.loc['80%'])
print(desc.loc['85.0%'])
print(desc.loc['90%'])
print(desc.loc['95%'])
print(desc.loc['100%'])

Expected Output

Indexes of percentiles should be either all whole numbers or either all decimal numbers.

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.5.2.final.0 python-bits: 64 OS: Linux OS-release: 4.15.0-50-generic machine: x86_64 processor: x86_64 byteorder: little LC_ALL: None LANG: en_US.UTF-8 LOCALE: en_US.UTF-8

pandas: 0.24.2
pytest: None
pip: 19.1.1
setuptools: 41.0.1
Cython: None
numpy: 1.16.3
scipy: 1.2.1
pyarrow: None
xarray: None
IPython: 7.5.0
sphinx: None
patsy: None
dateutil: 2.8.0
pytz: 2019.1
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 3.0.3
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: 0.7.3
lxml.etree: 3.5.0
bs4: 4.4.1
html5lib: 0.999
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.8
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None
gcsfs: None

[WillAyd]

Makes sense. If you'd like to take a look would certainly welcome a PR

[jkovacevic]

Would by my pleasure. I will follow up on this over this weekend.