pd.read_csv() reads a CSV (comma-separated values) file into a DataFrame, automatically inferring column names, data types, and handling most common formatting variations.

Syntax

pd.read_csv(filepath, sep=',', header=0, index_col=None, dtype=None)

Deep Dive Course

read_csv() is pandas' most commonly used I/O function, parsing a delimited text file into a fully structured DataFrame in one call — it automatically detects column headers from the first row by default, infers each column's dtype from its values, and handles many real-world quirks like quoted fields containing the delimiter, different line endings, and missing values represented as empty fields. The sep parameter changes the delimiter for non-comma-separated files, like tab-separated files, and index_col designates a specific column to use as the DataFrame's row labels instead of a default integer range.

1Understanding pd.read_csv()

💡

Pass parse_dates=['column_name'] to read_csv() to have it automatically convert a date column into proper datetime objects during loading, instead of reading it as plain text and converting separately afterward with pd.to_datetime().

editor.html

import pandas as pd

df = pd.read_csv("sales.csv")
print(df.head(2))

localhost:3000

2Practical Example

Here is a real-world application of pd.read_csv() showing how it is used in production Pandas code.

editor.html

import pandas as pd

df = pd.read_csv("sales.csv", usecols=["product", "price"], dtype={"price": float})
print(df.dtypes)

localhost:3000

3Best Practices

Follow these guidelines when working with pd.read_csv():

1. Specify dtype explicitly for columns pandas might misinterpret, like a numeric-looking ID column that should stay text, e.g. zip codes with leading zeros

2. Use parse_dates to convert date columns during loading, rather than as a separate follow-up step

3. Set index_col to a natural key column, like an ID or date, when that column is how you'll primarily look up or join rows

⚠️

Tip: Pass parse_dates=['column_name'] to read_csv() to have it automatically convert a date column into proper datetime objects during loading, instead of reading it as plain text and converting separately afterward with pd.to_datetime().

editor.html

import pandas as pd

df = pd.read_csv("sales.csv")
print(df.head(2))

localhost:3000

Examples

Example 01Basic Usage

import pandas as pd

df = pd.read_csv("sales.csv")
print(df.head(2))

Example 02Advanced Example

import pandas as pd

df = pd.read_csv("sales.csv", usecols=["product", "price"], dtype={"price": float})
print(df.dtypes)

Best Practices

Specify dtype explicitly for columns pandas might misinterpret, like a numeric-looking ID column that should stay text, e.g. zip codes with leading zeros
Use parse_dates to convert date columns during loading, rather than as a separate follow-up step
Set index_col to a natural key column, like an ID or date, when that column is how you'll primarily look up or join rows

Interview Question

Why might a numeric-looking ID column, like a US zip code with a leading zero, get corrupted by read_csv()'s default behavior?▼

Hint: Think about what read_csv() assumes when a column looks like it contains only digits.

By default, read_csv() infers each column's dtype from its values, and a column of digit-only strings gets automatically interpreted as an integer, which silently strips any leading zeros, since a leading zero has no meaning in a numeric value — a zip code like '02134' becomes the integer 2134. Passing an explicit string dtype for that column tells read_csv() to keep it as text instead of inferring a numeric type, preserving the leading zero and the column's original formatting.

Exercises

MediumPractice using pd.read_csv() in a real scenario.

View Solution

import pandas as pd

df = pd.read_csv("sales.csv")
print(df.head(2))

Frequently Asked Questions

Why might a numeric-looking ID column, like a US zip code with a leading zero, get corrupted by read_csv()'s default behavior?▼

Related Functions

pd-read-exceldf-to-csvpd-to-datetime