Pandas DataFrame Selecting and Indexing
import numpy as np
import pandas as pd
from pandas import Series, DataFrame
!ls *.csv
tmdb_5000_movies.csv?
imdb = pd.read_csv('tmdb_5000_movies.csv')
imdb.shape
(4803, 20)
imdb.head(2)
imdb.tail(2) # 查看最后兩行數(shù)據(jù)
imdb.columns # 查看column name
Index(['budget', 'genres', 'homepage', 'id', 'keywords', 'original_language',
'original_title', 'overview', 'popularity', 'production_companies',
'production_countries', 'release_date', 'revenue', 'runtime',
'spoken_languages', 'status', 'tagline', 'title', 'vote_average',
'vote_count'],
dtype='object')
sub_df = imdb[['budget','original_title','original_language','status']] # 生成新的dataframe
sub_df.shape
(4803, 4)
sub_df.iloc[10:12, :] # index location, 取10到11行，所有列
sub_df.iloc[:, 1:3].head(2) # 取1到3列洲劣，所有行的前兩行
sub_df.iloc[10:12,2:4] # 10和11行的2术浪、3列數(shù)據(jù)
temp_df = sub_df.iloc[10:20,0:2]
temp_df.loc[15:17,:] # 取15到16的label, 所有列蓝仲， 和Index沒(méi)關(guān)系
Reindexing Series 和 DataFrame
import numpy as np
import pandas as pd
from pandas import Series, DataFrame
s1 = Series([1, 2, 3, 4], index=['A', 'B', 'C', 'D'])
s1
A 1
B 2
C 3
D 4
dtype: int64
s1.reindex(index=['A','B', 'C', 'D', 'E']) # shift + tag 顯示提示幫助曙聂。 reindex生成一個(gè)新的Series對(duì)象
A 1.0
B 2.0
C 3.0
D 4.0
E NaN
dtype: float64
s1.reindex(index=['A','B', 'C', 'D', 'E'], fill_value=10) # 給源series對(duì)象不存在的列填充指定值
A 1
B 2
C 3
D 4
E 10
dtype: int64
s2 = Series(['A','B', 'C'], index=[1, 5, 10])
s2
1 A
5 B
10 C
dtype: object
s2.reindex(index=range(10), method='ffill') # frond fill
0 NaN
1 A
2 A
3 A
4 A
5 B
6 B
7 B
8 B
9 B
dtype: object
df1 = DataFrame(np.random.rand(25).reshape([5,5]), index=['A','B', 'D', 'E', 'F'], columns=['c1', 'c2','c3', 'c4', 'c5'])
df1
df1.reindex(index=['A', 'B', 'C', 'D','E', 'F'], columns=['c1', 'c2','c3', 'c4', 'c5', 'c6'])
s1
A 1
B 2
C 3
D 4
dtype: int64
s1.drop('A')
B 2
C 3
D 4
dtype: int64
df1.drop('A', axis=0) ## 刪除index：A行
df1.drop('c1', axis=1) ## 刪除column: c1行
多級(jí)Series
import numpy as np
import pandas as pd
from pandas import Series, DataFrame
s1 = Series(np.random.randn(6), index=[[1, 1, 1, 2, 2, 2], ['a', 'b' , 'c', 'a', 'b' , 'c']])
s1
1 a -1.896724
b 0.942334
c -0.020645
2 a -0.520883
b 1.740697
c 1.207718
dtype: float64
s1[1] # 取出1級(jí) 索引為1的值
print(type(s1)) # 返回一個(gè)series
<class 'pandas.core.series.Series'>
s1[:, 'a'] # 取出各級(jí)的'a'的值. 第一個(gè)冒號(hào)代指任意的1級(jí)菠红。
1 -1.896724
2 -0.520883
dtype: float64
- 多級(jí)Series和DataFrame轉(zhuǎn)換
df1 = s1.unstack() # 轉(zhuǎn)換series --> dataframe
df1
df2 = DataFrame([s1[1], s1[2]]) # 多級(jí)series的第一級(jí)豪直，分別當(dāng)成dataframe的元素金蜀，生成新的dataframe. 和s.unstack生成的一樣
- dataframe轉(zhuǎn)換為多級(jí)series
s2 = df1.unstack() # dataframe的unstack方法生成多級(jí)series
s3 = df1.T.unstack() # dataframe的unstack方法生成多級(jí)series吗伤。 T：會(huì)把column[1,2,1,2...]變成1級(jí)index吃靠，index數(shù)據(jù)[a, b]變成2級(jí)index
s2
a 1 -1.896724
2 -0.520883
b 1 0.942334
2 1.740697
c 1 -0.020645
2 1.207718
dtype: float64
s3
1 a -1.896724
b 0.942334
c -0.020645
2 a -0.520883
b 1.740697
c 1.207718
dtype: float64
多級(jí)DataFrame
df = DataFrame(np.arange(0,16).reshape([4,4]))
df # 具有1級(jí)column, 1級(jí)column的dataframe
df = DataFrame(np.arange(0,16).reshape([4,4]), index=[['a', 'a', 'b', 'b'], [1,1, 2, 2]]) # 具有多級(jí)的index
df
df = DataFrame(np.arange(0,16).reshape([4,4]), index=[['a', 'a', 'b', 'b'], [1,1, 2, 2]], columns=[['bj', 'bj', 'sh', 'gz'], [8, 9, 8, 8]]) # 具有多級(jí)的index,多級(jí)column
df
df['bj'] # 直接指定column訪問(wèn)
df['bj'][8] # 指定多級(jí)column訪問(wèn)
a 1 0
1 4
b 2 8
2 12
Name: 8, dtype: int64
