Pandas Dataframe行迭代并插入空行

Question

我是python中的新手。我有一个数据框，我将它写入excel。我要插入一个带子串的空行＆＃34; Total＆＃34;列＆＃34;放置ID＆＃34;值会更改并获取其他列的字段值的总和。

目前看来这样的数据。

http://prntscr.com/i0ndav

希望以这种方式转换它。

http://prntscr.com/i0nful

数据在这里

https://drive.google.com/open?id=15gef0zEReQb7qA7TsJSxIhYxLVVJLK4X

提前致谢。

Answer 1

这看起来怎么样？

# import pandas
import pandas as pd

# import data
df = pd.read_excel(r"C:\Users\IT045886\Downloads\Data.xlsx", skiprows = 12)

# convert str(floats) to actual floats
for col in df.loc[:, 'KM_Impressions':].columns:
    df[col] = pd.to_numeric(df[col].copy(), errors = 'coerce')

# limit column selection
df1 = df.loc[:, :'Deep Engagements'].copy()

# drop NaN and pivot
subs = df1.dropna().pivot_table(index = [
    'Placement ID',
    'Placement#',
    'Metric',
    'Day'
], aggfunc = 'sum')

# sum on first three levels
subs_t = subs.sum(level = [0,1,2])

# insert 'Total' in 'Day' column
subs_t.insert(0, 'Day', 'Total')

# set 'Day' in index
subs_t.set_index('Day', append = True, inplace = True)

# concat sums and totals
df2 = pd.concat([subs, subs_t]).sort_index().reset_index()

# merge original columns back in, reorder columns
final = df2.merge(df, how = 'outer')[df.columns]

final

pd.to_numeric()的文档 - ＆gt; https://pandas.pydata.org/pandas-docs/stable/generated/pandas.to_numeric.html

errors = 'coerce' # invalid parsing will be set as NaN