使用熊猫进行列到行转置的 Python 程序
原文:https://www.geesforgeks.org/python-program-for-column-to-row-transpose-use-pandas/
给定一个具有列“部门”和“名称”的输入文件，执行一个将列值转换为行的操作。名称包含属于由“部门”列标识的特定部门的管道分隔值
附加数据集: emp_data
示例:
Input:
dept, name
10, Vivek|John
20, Ritika|Shubham|Nitin
30, Vishakha|Ankit
Output:
dept, name
10, Vivek
10, John
20, Ritika
20, Shubham
20, Nitin
30, Vishakha
30, Ankit
方法一:蟒道
# Reading Data From the text
# file
data = pd.read_csv(r'GFG.txt')
# create new data frame with
# split value columns seperates
# data into three columns as per
# separator mentioned
new = data["name"].str.split("|",expand = True)
# making separate first name column
# from new data frame assign columnn
# values to dataframe new columns
# named as name*
data["Name1"] = new[0]
data["Name2"] = new[1]
data["Name3"] = new[2]
# Dropping old Name columns
data.drop(columns =["name"], inplace = True)
# create seperate dataframes with two
# columns id,name
d_name1 = data[['dept','Name1']]
d_name2 = data[['dept','Name2']]
d_name3 = data[['dept','Name3']]
# perform concat/unions operation for
# vertical merging of dataframes
union_df=pd.concat([d_name1,d_name2,d_name3],ignore_index=True)
union_df.fillna('',inplace=True)
# concatenate values of series into one
# series "name"
union_df['name'] = union_df['Name1'].astype(str)+union_df['Name2'].astype(str)+union_df['Name3'].astype(str)
# drop column names
union_df.drop(['Name1','Name2','Name3'],axis=1,inplace=True)
# drop rows having empty values
final_df=union_df[union_df['name']!='']
# sort the dataframe data by dept values
final_df.sort_values('dept')
输出:
注意:上述方法的缺点是当用|
方法二:探索熊猫
emp_df = pd.read_csv(r'GFG.txt')
# split column data on basis of seperator
# convert it into list using to_list
# stack method performs transpose operation
# to the data
emp_df1 = pd.DataFrame(emp_df.name.str.split('|').to_list(),
index = emp_df.dept).stack()
emp_df1 = emp_df1.reset_index([0, 'dept'])
emp_df1.columns =['Dept', 'Name']
emp_df1
输出:
方法三:熊猫方式:爆炸()
df = pd.read_csv(r'GFG.txt')
# seperate values using split()
# transpose is performed by explode
# function explode function overcomes
# the method1 shortcomings incase we
# have many columns we explode will do
# the task in no time and with no hassle
df1 = df.assign(name = df['name'].str.split('|')).explode('name')
df1
输出:
