pandas 1.0最新版本特性抢先看

pandas1.0最新版本特性抢先看

1. !pip3 install pandas==1.0.0rc0

1. Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple

2. Requirement already satisfied: pandas==1.0.0rc0 in /usr/local/lib/python3.7/site-packages (1.0.0rc0)

3. Requirement already satisfied: python-dateutil>=2.6.1 in /usr/local/lib/python3.7/site-packages (from pandas==1.0.0rc0) (2.8.0)

4. Requirement already satisfied: numpy>=1.13.3 in /usr/local/lib/python3.7/site-packages (from pandas==1.0.0rc0) (1.17.3)

5. Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.7/site-packages (from pandas==1.0.0rc0) (2019.3)

6. Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/site-packages (from python-dateutil>=2.6.1->pandas==1.0.0rc0) (1.12.0)

1. import pandas as pd

2. pd.__version__

1. '1.0.0rc0'

df.info()

dataframe最新的df.info()可以提供更加详尽的信息

1. import pandas as pd

2. 
   

3. df = pd.DataFrame({

4. 'A': [1,2,3],

5. 'B': ["goodbye", "cruel","world"],

6. 'C': [False, True, False]})

7. df.info()

1. <class 'pandas.core.frame.DataFrame'>

2. RangeIndex: 3 entries, 0 to 2

3. Data columns (total 3 columns):

4. # Column Non-Null Count Dtype

5. --- ------ -------------- -----

6. 0 A 3 non-null int64

7. 1 B 3 non-null object

8. 2 C 3 non-null bool

9. dtypes: bool(1), int64(1), object(1)

10. memory usage: 179.0+ bytes

df.to_markdown()

可以输出markdown表格，这个是我最喜欢的特性。因为公众号不支持pandas的输出结果，每次我都是截图粘贴。

1. print(df.to_markdown())

1. | | A | B | C |

2. |---:|----:|:--------|:------|

3. | 0 | 1 | goodbye | False |

4. | 1 | 2 | cruel | True |

5. | 2 | 3 | world | False |

新增bool和stirng两种数据类型

dataframe之前只支持object、int和float，其中object就是python对象。新版本新增bool(布尔型)和string(字符串)。

目前这个改变是试验性质，所以使用该api接口要注意。但pandas建议使用类型声明，未来pandas会改善这部分，可能增加更强大的正则匹配功能。

默认，pandas还会使用之前的object，除非你对字段数据类型声明为string或者bool。

1. import pandas as pd

2. 
   

3. B = pd.Series(["goodbye", "cruel", "world"], dtype="string")

4. C = pd.Series([False, True, False], dtype="bool")

5. df = pd.DataFrame({'B':B, 'C':C})

6. df.info()

1. <class 'pandas.core.frame.DataFrame'>

2. RangeIndex: 3 entries, 0 to 2

3. Data columns (total 2 columns):

4. # Column Non-Null Count Dtype

5. --- ------ -------------- -----

6. 0 B 3 non-null string

7. 1 C 3 non-null bool

8. dtypes: bool(1), string(1)

9. memory usage: 155.0 bytes

选择指定数据类型的字段

pandas的1.0更新，最有用的特性就是筛选指定数据类型的字段

1. df.select_dtypes("string")