Python从csv文件中读取数据及提取数据的方法

# 读取数据
import pandas as pd
data = pd.read_csv("data1.csv")
print(data)

# 读取数据
import pandas as pd
data = pd.read_csv("data1.csv",header=None)
print(data)

# 读取数据
import pandas as pd
data = pd.read_csv("data1.csv")
# print(data)
# ①获取所有列，并存入一个数组中
import numpy as np
data = np.array(data)
print(data) # 用户编号  性别  年龄(岁)  年收入(元)  是否购买
# [[15624510        1       19    19000        0]
#  [15810944        1       35    20000        0]
#  [15668575        2       26    43000        0]
#  [15603246        2       27    57000        0]
#  [  ...          ...      ...    ...       ...]]

# 读取数据
import pandas as pd
data = pd.read_csv("data1.csv")
print(data) # 用户编号  性别  年龄(岁)  年收入(元)  是否购买
# (1)获取第1列，并存入一个数组中
import numpy as np
col_1 = data["用户编号"]  #获取一列，用一维数据
data_1 = np.array(col_1)
print(data_1)
# [15624510 15810944 15668575 15603246 15804002 15728773 15598044 15694829
#  15600575 15727311 15570769 15606274 15746139 15704987 15628972 15697686
#  15733883 15617482 15704583 15621083 15649487 15736760 15714658 15599081
#  15705113 15631159 15792818 15633531 15744529]

# (2)获取第1，2列
col_12 = data[["用户编号","性别"]]  #获取两列，要用二维数据
data_12 = np.array(col_12)
print(data_12)
# [[15624510        1]
#  [15810944        1]
#  [15668575        2]
#  [15603246        2]
#  [  ...          ..]]

import pandas as pd
import numpy as np
data_1 = pd.read_csv("data1.csv",usecols=["用户编号"])
data_1 = np.array(data_1)
print(data_1)
# [[15624510]
#  [15810944]
#  [15668575]
#  [15603246]
#  [  ...   ]]

# (2)如获取第1,2列
data_12 = pd.read_csv("data1.csv",usecols=["用户编号","性别"])
data_12 = np.array(data_12)
print(data_12)
# [[15624510        1]
#  [15810944        1]
#  [15668575        2]
#  [15603246        2]
#  [  ...          ..]]

import pandas as pd
import numpy as np
data = pd.read_csv("data1.csv")
# （1）获取第1列
data_1 = data.iloc[:,0]
data_1 =np.array(data_1)
print(data_1)
# [15624510 15810944 15668575 15603246 15804002 15728773 15598044 15694829
#  15600575 15727311 15570769 15606274 15746139 15704987 15628972 15697686
#  15733883 15617482 15704583 15621083 15649487 15736760 15714658 15599081
#  15705113 15631159 15792818 15633531 15744529]

# （2）获取第1，2列
data_12 = data.iloc[:,0:2]
data_12 = np.array(data_12)
print(data_12)
# [[15624510        1]
#  [15810944        1]
#  [15668575        2]
#  [15603246        2]
#  [  ...          ..]]

# 获取最后两列
data_last = data.iloc[:,-2:]
data_last = np.array(data_last)
print(data_last)
# [[ 19000        0]
#  [ 20000        0]
#  [ 26    43000        0]
#  [ 27    57000        0]
#  [ ...    ...       ...]]