pandas 二

pandas更新中…

pandas版本：2.0.3

pandas两个比较重要的结构：Series和DataFrame

导包

import numpy as np
import pandas as pd
from pandas import Series, DataFrame

Series 的reindex操作

s1 = pd.Series([1, 2, 3], index=["a", "b", "c"])
print(s1)
# a    1
# b    2
# c    3
# dtype: int64
print(s1.reindex(["a", "b", "c", "d"]))
# a    1.0
# b    2.0
# c    3.0
# d    NaN
# dtype: float64
print(s1.reindex(["a", "b", "c", "d"], fill_value=666))
# a      1
# b      2
# c      3
# d    666
# dtype: int64

s2 = pd.Series(["A", "B", "C"], index=[1, 3, 5])
print(s2)
# 增加新的index索引
# method="ffill"方式[1,3)填充index=1的值，[3,5)填充index=3的值，[5,7)填充index=5的值
print(s2.reindex(range(7), method="ffill"))

reindex 也可以用来截取数据

print(s1.reindex(index=["a", "b"]))
# a    1
# b    2
# dtype: int64
# 等价
print(s1.drop("a"))
# b    2
# c    3
# dtype: int64

DataFrame的reindex操作

df1 = DataFrame(
    np.random.rand(25).reshape(5, 5),
    index=["A", "B", "D", "E", "F"],
    columns=["c1", "c2", "c3", "c4", "c5"],
)
print(df1)
print(df1.reindex(index=["A", "B", "C", "D", "E", "F"]))
#          c1        c2        c3        c4        c5
# A  0.205651  0.094378  0.548683  0.104871  0.497516
# B  0.086842  0.842261  0.912813  0.661971  0.849072
# C       NaN       NaN       NaN       NaN       NaN
# D  0.570752  0.585542  0.606429  0.713678  0.013199
# E  0.650421  0.567925  0.303060  0.234874  0.911466
# F  0.011784  0.304507  0.040120  0.064653  0.209229
print(df1.reindex(columns=["c1", "c2", "c3", "c4", "c5","c6"])) 
# !!可以看出reindex方法返回新的DataFrame
#          c1        c2        c3        c4        c5  c6
# A  0.279706  0.179977  0.744959  0.637003  0.574922 NaN
# B  0.179284  0.159424  0.857709  0.302644  0.051491 NaN
# D  0.571989  0.893290  0.337798  0.385577  0.098375 NaN
# E  0.128973  0.367170  0.556358  0.303979  0.628117 NaN
# F  0.358955  0.851620  0.789938  0.720290  0.169392 NaN

# 增加行列
print(df1.reindex(
    index=["A", "B", "C", "D", "E", "F"], 
    columns=["c1", "c2", "c3", "c4", "c5", "c6"]
))
#          c1        c2        c3        c4        c5  c6
# A  0.600788  0.478370  0.842493  0.645647  0.927288 NaN
# B  0.466357  0.523482  0.792704  0.004789  0.319052 NaN
# C       NaN       NaN       NaN       NaN       NaN NaN
# D  0.677526  0.039438  0.846586  0.875791  0.354747 NaN
# E  0.577882  0.779355  0.089818  0.958951  0.697171 NaN
# F  0.130449  0.671984  0.922430  0.902640  0.100657 NaN

reindex也可以用来截取数据

print(df1.reindex(index=["A","B"]))
#          c1        c2        c3        c4        c5
# A  0.600788  0.478370  0.842493  0.645647  0.927288
# B  0.466357  0.523482  0.792704  0.004789  0.319052

print(df1.drop("A", axis=0)) # 行删除
#          c1        c2        c3        c4        c5
# B  0.466357  0.523482  0.792704  0.004789  0.319052
# D  0.677526  0.039438  0.846586  0.875791  0.354747
# E  0.577882  0.779355  0.089818  0.958951  0.697171
# F  0.130449  0.671984  0.922430  0.902640  0.100657

print(df1.drop("c1", axis=1)) # 列删除
#          c2        c3        c4        c5
# A  0.478370  0.842493  0.645647  0.927288
# B  0.523482  0.792704  0.004789  0.319052
# D  0.039438  0.846586  0.875791  0.354747
# E  0.779355  0.089818  0.958951  0.697171
# F  0.671984  0.922430  0.902640  0.100657

Nan操作

Series中的nan

s1 = pd.Series([1, 2, np.nan, 3, 4], index=["A", "B", "C", "D", "E"])
print(s1)
# A    1.0
# B    2.0
# C    NaN
# D    3.0
# E    4.0
# dtype: float64

print(s1.isnull()) # nan为True
# A    False
# B    False
# C     True
# D    False
# E    False
# dtype: bool

s1.dropna() # 删除nan
# A    1.0
# B    2.0
# D    3.0
# E    4.0
# dtype: float64

DataFrame的nan

dframe = DataFrame(
    [[1, 2, 3], [np.nan, 5, 6], [7, np.nan, 8], [np.nan, np.nan, np.nan]]
)
print(dframe)
#      0    1    2
# 0  1.0  2.0  3.0
# 1  NaN  5.0  6.0
# 2  7.0  NaN  8.0
# 3  NaN  NaN  NaN

print(dframe.isnull())
#        0      1      2
# 0  False  False  False
# 1   True  False  False
# 2  False   True  False
# 3   True   True   True

过滤和填充nan

# axis=0, how="any" axis删除行的nan值，how=方式为包含就删除
# !!上述两个参数默认值就是0和any
# axis和numpy的axis参数不一样
# 等价 print(dframe.dropna())
df1 = dframe.dropna(axis=0, how="any")
print(df1)
#      0    1    2
# 0  1.0  2.0  3.0

df2 = dframe.dropna(thresh=2)  # 行大于2个的nan将被删除
print(df2)
#      0    1    2
# 0  1.0  2.0  3.0
# 1  NaN  5.0  6.0
# 2  7.0  NaN  8.0

df3 = dframe.fillna(value=99) # 填充nan
print(df3)
#       0     1     2
# 0   1.0   2.0   3.0
# 1  99.0   5.0   6.0
# 2   7.0  99.0   8.0
# 3  99.0  99.0  99.0

map操作

df1 = DataFrame({"城市":["北京","上海","广州"],
                  "人口":[1000,2000,1500]})
print(df1)
#    城市    人口
# 0  北京  1000
# 1  上海  2000
# 2  广州  1500

# gdp_map = {"北京": 1, "上海": 2, "广州": 3}
gdp_map = {"北京": 1,  "广州": 3} # 缺省值为nan
df1["GDP"] = df1["城市"].map(gdp_map)
print(df1)
#    城市    人口  GDP
# 0  北京  1000  1.0
# 1  上海  2000  NaN
# 2  广州  1500  3.0