240 lines
5.6 KiB
Markdown
240 lines
5.6 KiB
Markdown
|
# 将 Pandas 数据帧转换为 Numpy 数组[分步]
|
|||
|
|
|
|||
|
|
> 原文：<https://www.askpython.com/python-modules/numpy/pandas-dataframe-to-numpy-array>
|
|||
|
|
|
|||
|
|
读者你好！在本文中，我们将了解什么是数据帧，以及如何将 Pandas 数据帧转换为 Numpy 数组，反之亦然。让我们开始吧:
|
|||
|
|
|
|||
|
|
## 介绍
|
|||
|
|
|
|||
|
|
Python 中的数据框是一种二维表格数据结构，由定义数据框不同特征的行和列组成。
|
|||
|
|
|
|||
|
|
我们可以使用 [Pandas 库](https://www.askpython.com/python-modules/pandas/python-pandas-module-tutorial)创建一个数据框，或者导入一个已经构建好的数据框(。csv 文件)并处理它。你可以使用 [pip 命令](https://www.askpython.com/python-modules/python-pip)安装 Pandas。
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
pip install pandas
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
上面写的代码安装了 pandas，我们都准备好使用 Pandas 库的不同功能了。同样，我们将安装 [numpy 库](https://www.askpython.com/python-modules/numpy/python-numpy-arrays):
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
pip install numpy
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
首先，我们将看到这个表格结构(pandas 数据帧)到 numpy 数组的转换。
|
|||
|
|
|
|||
|
|
* * *
|
|||
|
|
|
|||
|
|
### 1.将 Pandas 数据帧转换为 Numpy 数组
|
|||
|
|
|
|||
|
|
我们可以通过使用 *dataframe.to_numpy()* 方法来做到这一点。这将给定的 Pandas 数据帧转换成 Numpy 数组。
|
|||
|
|
|
|||
|
|
* 让我们创建两个数据框，我们将在本教程中使用。
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
#importing pandas
|
|||
|
|
import pandas as pd
|
|||
|
|
|
|||
|
|
#creating dataframes
|
|||
|
|
student_data = {"Name": ['Alice', 'Sam', 'Kevin', 'Max', 'Tom'],
|
|||
|
|
"exam_no": [201, 202, 203, 204, 205],
|
|||
|
|
"Result": ['Pass', 'Pass', 'Fail', 'Pass', 'Fail']}
|
|||
|
|
|
|||
|
|
set_of_numbers = {"Numbers": ['134', '273', '325','69.21','965']}
|
|||
|
|
|
|||
|
|
print("This is our first dataset :")
|
|||
|
|
student_dataframe = pd.DataFrame(student_data)
|
|||
|
|
print("\n",student_dataframe)
|
|||
|
|
|
|||
|
|
print("\nThis is our second dataset :")
|
|||
|
|
numbers_dataframe = pd.DataFrame(set_of_numbers)
|
|||
|
|
print("\n",numbers_dataframe)
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
* 我们创建了两个数据帧:**学生数据**和**数字集合**。我们的数据框看起来像这样:
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
* 现在，在将 Pandas 数据帧转换为 Numpy 数组之前，让我们来看看类型:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
print(type(student_dataframe))
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
print(type(numbers_dataframe))
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
上面两条语句的输出是相同的。即，
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
<class 'pandas.core.frame.DataFrame'>
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
* 要将这个 Pandas 数据帧转换成 Numpy 数组，运行下面给出的代码
|
|||
|
|
|
|||
|
|
将学生数据转换为
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
student_array = student_dataframe.to_numpy()
|
|||
|
|
print(student_array)
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
输出:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
[['Alice' 201 'Pass']
|
|||
|
|
['Sam' 202 'Pass']
|
|||
|
|
['Kevin' 203 'Fail']
|
|||
|
|
['Max' 204 'Pass']
|
|||
|
|
['Tom' 205 'Fail']]
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
对于第二数据帧(组数)
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
numbers_array = numbers_dataframe.to_numpy()
|
|||
|
|
print(numbers_array)
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
输出:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
[['134']
|
|||
|
|
['273']
|
|||
|
|
['325']
|
|||
|
|
['69.21']
|
|||
|
|
['965']]
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
* 我们还可以检查两个数组的数据类型:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
print(type(student_array))
|
|||
|
|
print(type(numbers_array))
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
输出:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
<class 'numpy.ndarray'>
|
|||
|
|
<class 'numpy.ndarray'>
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
因此，我们可以清楚地看到，我们只用了几个步骤就将 Pandas 数据帧转换为 Numpy 数组。这是处理数据帧及其转换的最简单方法。
|
|||
|
|
|
|||
|
|
* 此外，我们还可以更改数据框中列的数据类型。考虑到我们的第二个数据帧，它由一些整数值和一些浮点值组成，让我们尝试将它们都改为浮点型。
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
print(numbers_dataframe.to_numpy(dtype ='float64'))
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
输出:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
[[134\. ]
|
|||
|
|
[273\. ]
|
|||
|
|
[325\. ]
|
|||
|
|
[ 69.21]
|
|||
|
|
[965\. ]]
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
* * *
|
|||
|
|
|
|||
|
|
### 2.将 Numpy 数组转换为 Pandas 数据帧
|
|||
|
|
|
|||
|
|
现在您已经理解了 Pandas 数据帧到 Numpy 数组的转换，我们可能需要将数据转换回 Numpy 数组。让我们看看如何做到这一点:
|
|||
|
|
|
|||
|
|
* 首先，定义一个 [numpy 数组](https://www.askpython.com/python-modules/numpy/python-numpy-arrays)。然后使用 pandas 执行转换。熊猫图书馆的 DataFrame()函数。
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
#importing pandas and numpy
|
|||
|
|
import pandas as pd
|
|||
|
|
import numpy as np
|
|||
|
|
|
|||
|
|
#defining numpy array
|
|||
|
|
arr1 = np.array([[1,6,4,5], [3,7,2,4], [9,5,3,7]])
|
|||
|
|
print("Numpy array : ")
|
|||
|
|
print(arr1)
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
我们的数组是这样的:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
Numpy array :
|
|||
|
|
[[1 6 4 5]
|
|||
|
|
[3 7 2 4]
|
|||
|
|
[9 5 3 7]]
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
* 现在，把它转换成熊猫数据帧:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
#converting array to dataframe
|
|||
|
|
df = pd.DataFrame(arr1)
|
|||
|
|
print("\npandas dataframe :")
|
|||
|
|
df
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
转换后的数据帧为:
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
* 检查数据帧的类型:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
type(df)
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
输出:
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
pandas.core.frame.DataFrame
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
* 我们还可以为数据帧的行和列指定自己的标题。可以使用关键字 *index* 给出行的标题，对于列，我们使用关键字 *columns* 。
|
|||
|
|
|
|||
|
|
```py
|
|||
|
|
#converting and providing headers
|
|||
|
|
df = pd.DataFrame(arr1, index = ["1","2","3"], columns = ["A","B","C","D" ])
|
|||
|
|
print("\npandas dataframe :")
|
|||
|
|
df
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
这将使我们的数据框看起来像这样:
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
## 结论
|
|||
|
|
|
|||
|
|
就这样，我们到了这篇文章的结尾。在这篇文章中，你明白了
|
|||
|
|
|
|||
|
|
* 熊猫数据帧和数组的基本知识
|
|||
|
|
* 如何将 pandas 数据帧转换为 numpy 数组
|
|||
|
|
* 如何将 numpy 数组转换为 pandas 数据帧
|
|||
|
|
|
|||
|
|
我希望这篇文章对你有用。谢谢大家！🙂
|
|||
|
|
|
|||
|
|
***参考文献—***
|
|||
|
|
|
|||
|
|
* *[https://pandas . pydata . org/pandas-docs/stable/reference/API/pandas。data frame . to _ numpy . html](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.to_numpy.html)*
|
|||
|
|
* [*https://stack overflow . com/questions/13187778/convert-pandas-data frame-to-numpy-array/*](https://stackoverflow.com/questions/13187778/convert-pandas-dataframe-to-numpy-array/)
|