# dtype

*noun*

say it: **“DEE-type”** **/ˈdiˌtaɪp/**

In pandas and NumPy, the kind of value a column or array holds, such as whole numbers, decimal numbers, text or dates.

## What is a **dtype** in pandas?

A **dtype**, short for data type, is the kind of value every row of a pandas column holds: `int64` for whole
numbers, `float64` for numbers with decimals, `str` for text and `datetime64` for dates. It decides what the column
can do: a `str` column of weights can't be averaged until it becomes numbers.

```python
import pandas as pd
weights = pd.Series(["12 lbs", "5.5", "about 15"])
print(weights.dtype)                       # str: text, not numbers
print(pd.Series([21, 25, 18]).dtype)       # int64
```

## In a sentence

- “The weight column's **dtype** was str, so its mean couldn't be worked out.”
- “info() lists the **dtype** of every column.”

Level: Journeyman · [More Python terms at this level](https://www.ask-codey.com/glossary?category=python&level=3)

Updated: November 19, 2026

Source: https://www.ask-codey.com/glossary/dtype
