Python 中的 Pandas 插入方法
Suraj Joshi
2023年1月30日
本教程解釋瞭如何使用 insert() 方法在 Pandas DataFrame 中插入一列。
import pandas as pd
countries_df = pd.DataFrame(
{
"Country": ["Nepal", "Switzerland", "Germany", "Canada"],
"Continent": ["Asia", "Europe", "Europe", "North America"],
"Primary Language": ["Nepali", "French", "German", "English"],
}
)
print("Countries DataFrame:")
print(countries_df, "\n")
輸出：
Countries DataFrame:
Country Continent Primary Language
0 Nepal Asia Nepali
1 Switzerland Europe French
2 Germany Europe German
3 Canada North America English
我們將使用上例中所示的 countries_df DataFrame 來解釋如何使用 insert() 方法在 Pandas DataFrame 中插入一列。
pandas.DataFrame.insert() 方法
語法
DataFrame.insert(loc, column, value, allow_duplicates=False)
它將名為 column 的列插入到 DataFrame 中，其值由 value 指定，位於 loc 位置。
使用 insert() 方法插入對所有行具有相同值的列
import pandas as pd
countries_df = pd.DataFrame(
{
"Country": ["Nepal", "Switzerland", "Germany", "Canada"],
"Continent": ["Asia", "Europe", "Europe", "North America"],
"Primary Language": ["Nepali", "French", "German", "English"],
}
)
print("Countries DataFrame:")
print(countries_df, "\n")
countries_df.insert(3, "Capital", "Unknown")
print("Countries DataFrame after inserting Capital column:")
print(countries_df)
輸出：
Countries DataFrame:
Country Continent Primary Language
0 Nepal Asia Nepali
1 Switzerland Europe French
2 Germany Europe German
3 Canada North America English
Countries DataFrame after inserting Capital column:
Country Continent Primary Language Capital
0 Nepal Asia Nepali Unknown
1 Switzerland Europe French Unknown
2 Germany Europe German Unknown
3 Canada North America English Unknown
它在 countries_df DataFrame 的 3 索引位置插入 ·Capital·列，所有行的 ·Capital·列值均設定為 Unknown。
該位置從 0 開始，因此 3 位置指的是 DataFrame 中的 4 列。
在 DataFrame 中插入一列，指定每行的值
如果我們想使用 insert() 方法為要插入的列指定每一行的值，我們可以在 insert() 方法中傳遞一個值列表作為 value 引數。
import pandas as pd
countries_df = pd.DataFrame(
{
"Country": ["Nepal", "Switzerland", "Germany", "Canada"],
"Continent": ["Asia", "Europe", "Europe", "North America"],
"Primary Language": ["Nepali", "French", "German", "English"],
}
)
print("Countries DataFrame:")
print(countries_df, "\n")
capitals = ["Kathmandu", "Zurich", "Berlin", "Ottawa"]
countries_df.insert(2, "Capital", capitals)
print("Countries DataFrame after inserting Capital column:")
print(countries_df)
輸出：
Countries DataFrame:
Country Continent Primary Language
0 Nepal Asia Nepali
1 Switzerland Europe French
2 Germany Europe German
3 Canada North America English
Countries DataFrame after inserting Capital column:
Country Continent Capital Primary Language
0 Nepal Asia Kathmandu Nepali
1 Switzerland Europe Zurich French
2 Germany Europe Berlin German
3 Canada North America Ottawa English
它在 DataFrame countries_df 中的索引 2 插入了列 Capital，併為 DataFrame 中的 Capital 列指定了每一行的值。
在 insert() 方法中設定 allow_duplicates = True 來新增已經存在的列
import pandas as pd
countries_df = pd.DataFrame(
{
"Country": ["Nepal", "Switzerland", "Germany", "Canada"],
"Continent": ["Asia", "Europe", "Europe", "North America"],
"Primary Language": ["Nepali", "French", "German", "English"],
"Capital": ["Kathmandu", "Zurich", "Berlin", "Ottawa"],
}
)
print("Countries DataFrame:")
print(countries_df, "\n")
capitals = ["Kathmandu", "Zurich", "Berlin", "Ottawa"]
countries_df.insert(4, "Capital", capitals, allow_duplicates=True)
print("Countries DataFrame after inserting Capital column:")
print(countries_df)
輸出：
Countries DataFrame:
Country Continent Primary Language Capital
0 Nepal Asia Nepali Kathmandu
1 Switzerland Europe French Zurich
2 Germany Europe German Berlin
3 Canada North America English Ottawa
Countries DataFrame after inserting Capital column:
Country Continent Primary Language Capital Capital
0 Nepal Asia Nepali Kathmandu Kathmandu
1 Switzerland Europe French Zurich Zurich
2 Germany Europe German Berlin Berlin
3 Canada North America English Ottawa Ottawa
它將列 Capital 新增到 countries_df DataFrame 中，儘管 countries_df DataFrame 中已經存在 Capital 列。
如果我們嘗試插入已經存在於 DataFrame 中的列，而沒有在 insert() 方法中設定 allow_duplicates = True，它就會向我們丟擲一個錯誤資訊：ValueError: cannot insert column, already exists.。
作者： Suraj Joshi
Suraj Joshi is a backend software engineer at Matrice.ai.
LinkedIn