Series Attributes

Before you analyse a column, you want quick facts about it: how many values, what type, are there missing values, are the labels unique? Attributes answer these questions instantly, without changing the data.

Definition

An attribute is a property stored on an object that you read without parentheses (s.shape), while a method is a function you call with parentheses (s.count()). Series attributes describe the Series' structure and contents — its size, type, name and labels.

• Without parentheses — s.dtype ✅, s.dtype() ❌ (TypeError: ... object is not callable).
• Describe, not change — reading an attribute never modifies the Series (a few, like name and index, can be assigned).

Syntax

s.shape        s.size        s.ndim        s.dtype
s.name         s.index       s.values      s.is_unique
s.hasnans      s.empty       s.nbytes      s.is_monotonic_increasing
s.info()       len(s)        s.count()     # helpers (methods / built-ins)

Example — all the basic attributes

import pandas as pd

runs = pd.Series([45, 78, None, 101, 33],
                 index=["Rohit", "Virat", "Gill", "Pant", "Hardik"],
                 name="runs")
print(runs)
print("shape   :", runs.shape)
print("size    :", runs.size)
print("len     :", len(runs))
print("ndim    :", runs.ndim)
print("dtype   :", runs.dtype)
print("name    :", runs.name)
print("is_unique:", runs.is_unique)
print("hasnans :", runs.hasnans)
print("empty   :", runs.empty)
print("count() :", runs.count())
print("nbytes  :", runs.nbytes)
print("index   :", runs.index)

Rohit      45.0
Virat      78.0
Gill        NaN
Pant      101.0
Hardik     33.0
Name: runs, dtype: float64
shape   : (5,)
size    : 5
len     : 5
ndim    : 1
dtype   : float64
name    : runs
is_unique: True
hasnans : True
empty   : False
count() : 4
nbytes  : 40
index   : Index(['Rohit', 'Virat', 'Gill', 'Pant', 'Hardik'], dtype='str')

How it works:

• shape is (5,) — a tuple with one number because a Series has one dimension.
• size and len count all 5 slots, while count() counts only the 4 non-missing values. size - count() = number of missing values.
• Gill's None became NaN, so dtype is float64 and hasnans is True.
• nbytes is 40 because 5 float64 values × 8 bytes each = 40.
• The index holds text labels, so it shows dtype='str'.

Example — is_unique checks values, not labels

teams = pd.Series(["MI", "CSK", "MI", "RCB"])
print(teams.is_unique, teams.index.is_unique, teams.dtype)

False True str

"MI" appears twice, so the values are not unique. The index 0–3 is unique — check it with s.index.is_unique.

Example — info()

runs.info()

<class 'pandas.Series'>
Index: 5 entries, Rohit to Hardik
Series name: runs
Non-Null Count  Dtype
--------------  -----
4 non-null      float64
dtypes: float64(1)
memory usage: 80.0+ bytes

How it works: info() prints (and returns None) a summary: the class, the index range (Rohit to Hardik), the name, how many values are not missing, the dtype, and memory used. With a default index it would say RangeIndex: 5 entries, 0 to 4. The + in 80.0+ means "at least" — text labels can use more memory than shown.

Example — setting name and renaming

r = runs.rename("runs_2024")   # returns a NEW Series with a new name
print(r.name, runs.name)
runs.name = "score"            # assign directly
print(runs.name)

runs_2024 runs
score

Common mistakes

• ❌ s.shape() or s.dtype(). ✅ Attributes have no parentheses: s.shape, s.dtype.
• ❌ Using size to count real data. ✅ size includes NaN; use count() for non-missing values.
• ❌ Thinking is_unique checks the labels. ✅ It checks the values; use s.index.is_unique for labels.
• ❌ x = s.info() and then using x. ✅ info() prints and returns None.

Interview tip

Q: How do you count missing values in a Series? A: s.isna().sum() — or s.size - s.count(). Use s.hasnans for a quick yes/no check.