# 查看索引数据 ```python import pandas as pd import numpy as np ``` ## 创建数据 ```python s = pd.Series([1,2,3,np.nan,5]) date = pd.date_range('20170101',periods=7) s=pd.DataFrame(np.random.rand(7,4),index=date,columns=['A','B','C','D']) s ''' A B C D 2017-01-01 0.756081 0.273180 0.420455 0.632778 2017-01-02 0.823846 0.020843 0.365796 0.070497 2017-01-03 0.746543 0.182495 0.183828 0.290980 2017-01-04 0.480208 0.934172 0.558363 0.443208 2017-01-05 0.198488 0.566854 0.064392 0.106457 2017-01-06 0.639944 0.017557 0.744136 0.133560 2017-01-07 0.449300 0.792805 0.098290 0.716538 ''' ``` ## 查看 ### 获取列 s\['A'] 和 s\[\['A']] 、s\[\['A','C']] ```python # 获得数据，按照列 名称 print(s['A'],'\n',s['A'].shape) ''' 2017-01-01 0.362362 2017-01-02 0.261357 2017-01-03 0.567673 2017-01-04 0.971732 2017-01-05 0.163707 2017-01-06 0.505815 2017-01-07 0.739000 Freq: D, Name: A, dtype: float64 (7,) ''' ``` ```python print(s[['A']],'\n',s[['A']].shape) ''' A 2017-01-01 0.362362 2017-01-02 0.261357 2017-01-03 0.567673 2017-01-04 0.971732 2017-01-05 0.163707 2017-01-06 0.505815 2017-01-07 0.739000 (7, 1) ''' ``` 获得数据，按照列 名称 多个 ```python print(s[['A','C']],s[['A','C']].shape) A C 2017-01-01 0.325836 0.286616 2017-01-02 0.915091 0.907306 2017-01-03 0.540354 0.533346 2017-01-04 0.359208 0.466159 2017-01-05 0.513239 0.329017 2017-01-06 0.793126 0.382535 2017-01-07 0.444599 0.583308 (7, 2) ``` 也可以直接加名称，不过这样一次只能取一个 ```python # 等价 s['A'] s.A ''' 2017-01-01 0.756081 2017-01-02 0.823846 2017-01-03 0.746543 2017-01-04 0.480208 2017-01-05 0.198488 2017-01-06 0.639944 2017-01-07 0.449300 Freq: D, Name: A, dtype: float64 ''' ``` ### 获取行 与列一样，也有数组形式 ```python # 前2行，**列名也算一行哦** print(s[:2]) ''' A B C D 2017-01-01 0.756081 0.273180 0.420455 0.632778 2017-01-02 0.823846 0.020843 0.365796 0.070497''' ``` ```python # 获得数据，按照行 名称 print(s.loc['2017-01-02'],s.loc['2017-01-02'].shape) ''' A 0.823846 B 0.020843 C 0.365796 D 0.070497 Name: 2017-01-02 00:00:00, dtype: float64 (4,) ''' ``` 同时选择 行、列 **loc、iloc** ```python # 依据 行、列切数据；（前面只是是个：范围区间，灵活位置数组选择只能用iloc） print(s.loc['2017-01-02':'2017-01-04',['A','D']]) ''' A D 2017-01-02 0.823846 0.070497 2017-01-03 0.746543 0.290980 2017-01-04 0.480208 0.443208 ''' ``` ```python # 获得数据，按照：行序号，列序号 s.iloc[[1,2,3],[0,3]] ''' A D 2017-01-02 0.823846 0.070497 2017-01-03 0.746543 0.290980 2017-01-04 0.480208 0.443208 ''' ``` ### 查看头尾 ```python print(df2.head(n=2),'\n',df2.tail(n=2)) ``` ### 查看索引、列、numpy值 ```python print(df2.index) print(df2.columns) print(df2.values) ''' Int64Index([0, 1, 2, 3], dtype='int64') Index(['A', 'B', 'C', 'D', 'E', 'F'], dtype='object') [[1.0 Timestamp('2017-01-02 00:00:00') 11 3 'test' 'foo'] [1.0 Timestamp('2017-01-02 00:00:00') 22 3 'train' 'foo'] [1.0 Timestamp('2017-01-02 00:00:00') 32 3 'test' 'foo'] [1.0 Timestamp('2017-01-02 00:00:00') 44 3 'train' 'foo']] ''' ``` ### 快速统计摘要 计算数值类型数据的，均值，方差等 ```python # B，E，F不是数值 df2.describe() ''' A C D count 4.0 4.000000 4.0 mean 1.0 27.250000 3.0 std 0.0 14.080128 0.0 min 1.0 11.000000 3.0 25% 1.0 19.250000 3.0 50% 1.0 27.000000 3.0 75% 1.0 35.000000 3.0 max 1.0 44.000000 3.0 ''' ``` ### 转置数据 ```python ''' A B C D E F 0 1.0 2017-01-02 11 3 test foo 1 1.0 2017-01-02 22 3 train foo 2 1.0 2017-01-02 32 3 test foo 3 1.0 2017-01-02 44 3 train foo ''' df2.columns ''' Index(['A', 'B', 'C', 'D', 'E', 'F'], dtype='object') ''' df2.T.columns ''' Int64Index([0, 1, 2, 3], dtype='int64') ''' df2.T ''' 0 1 2 3 A 1 1 1 1 B 2017-01-02 00:00:00 2017-01-02 00:00:00 2017-01-02 00:00:00 2017-01-02 00:00:00 C 11 22 32 44 D 3 3 3 3 E test train test train F foo foo foo foo ''' ``` ### 轴排序 ```python ''' A B C D E F 0 1.0 2017-01-02 11 3 test foo 1 1.0 2017-01-02 22 3 train foo 2 1.0 2017-01-02 32 3 test foo 3 1.0 2017-01-02 44 3 train foo ''' # axis=0，对index排序， 根据columns:C 下降排序，默认inplace False,不对原值替换 df2.sort_index(axis=0, by='C',ascending=False,inplace=True) # 重新排序对 df df2 ''' A B C D E F 3 1.0 2017-01-02 44 3 train foo 2 1.0 2017-01-02 32 3 test foo 1 1.0 2017-01-02 22 3 train foo 0 1.0 2017-01-02 11 3 test foo ''' ``` ### 布尔选择 ```python # Series类型，满足条件的索引 index=df2['C']>25 print(index) df2[df2['C']>25] ''' 3 True 2 True 1 False 0 False Name: C, dtype: bool A B C D E F 3 1.0 2017-01-02 44 3 train foo 2 1.0 2017-01-02 32 3 test foo ''' ``` ### isin ```python # E 列中只要存在 ['abc','train',666]中就行的匹配 df2[df2['E'].isin(['abc','train',666])] ''' A B C D E F 3 1.0 2017-01-02 44 3 train foo 1 1.0 2017-01-02 22 3 train foo ''' ``` ## 修改数据类型 ```python ''' A B C D E F 0 1.0 2017-01-02 11.0 3 test foo 1 1.0 2017-01-02 22.0 3 train foo 2 1.0 2017-01-02 32.0 3 test foo 3 1.0 2017-01-02 44.0 3 train foo ''' df2['C']=df2['C'].astype(np.int) ''' A B C D E F 0 1.0 2017-01-02 11 3 test foo 1 1.0 2017-01-02 22 3 train foo 2 1.0 2017-01-02 32 3 test foo 3 1.0 2017-01-02 44 3 train foo ''' ``` --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://im-qianuxn.gitbook.io/pytorch/ji-suan-ji/numpy-pandas-matplotlib/pandas/suo-yin-shu-ju.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.