Selecting and filtering data

Boolean filtering is used to select rows by condition in a DataFrame.

Filtering with a condition

The boolean mask from NumPy works in pandas too:

genc = df[df["yas"] < 30]
zengin_genc = df[(df["yas"] < 30) & (df["maas"] > 4000)]
print(len(genc))

Parentheses are required when combining conditions with & and |; Python's and/or keywords don't work here.

loc and iloc

loc selects by label, iloc by position; both use the [row, column] form:

df.loc[df["yas"] > 30, "ad"]   # ad column of the matching rows
df.iloc[0]                     # first row
df.iloc[:2, 1]                 # second column of the first two rows

Doing the filter and the column selection in one step with loc is readable and also avoids the "SettingWithCopy" warning.

Exercise Filter and select

Select the names of people older than 25 with a salary above 4500 using loc, and print them as a list with tolist().

Show example solution

import pandas as pd

df = pd.DataFrame({
    "ad": ["Ali", "Ece", "Can", "Su"],
    "yas": [34, 28, 45, 22],
    "maas": [5000, 4200, 7100, 3000],
})

secilen = df.loc[(df["yas"] > 25) & (df["maas"] > 4500), "ad"]
print(secilen.tolist())

Test yourself

Why does <code>df[df["yas"] > 30 and df["maas"] > 0]</code> error?
- 30 must be text
- It doesn't error
- and needs a single truth value of a Series; & and parentheses are needed
- The column doesn't exist
Show answer

and needs a single truth value of a Series; & and parentheses are needed — For element-wise "and" you write (condition1) & (condition2).
Select the first row by position.
```
ilk = df.____[0]
```
Show answer
```
ilk = df.iloc[0]
```
Match the selectors.

loc

Selection by label
iloc

Selection by position
&

Element-wise and
|

Element-wise or

Next lesson: Viewing first rows

We Are Enhancing Your Experience

Cookie Preferences

Strictly Necessary Cookies

Analytical Cookies