[Python] Fast Retrieval and Update of Specific Values in Pandas (at / iat)

2026年1月5日

In Pandas DataFrames, when you want to pinpoint and retrieve or update a single value (scalar value) by specifying a specific “row” and “column,” the at and iat accessors are provided as faster alternatives to loc and iloc.

at: Specified by row label (index name) and column name.
iat: Specified by row number and column number (integers).

This article explains how to retrieve and update values using these accessors.

Implementation Sample Code

Below is code that retrieves and modifies information for specific products using cafe menu data (price and calories).

import pandas as pd

def access_scalar_values():
    """
    Function to manipulate single values in a DataFrame using at and iat
    """
    
    # 1. Data Preparation
    # DataFrame with drink names as index and price/calories as columns
    menu_items = ["Espresso", "Latte", "Mocha", "Cappuccino", "Americano"]
    menu_data = {
        "Price": [350, 450, 500, 450, 380],
        "Calories": [10, 150, 280, 120, 15]
    }
    
    df = pd.DataFrame(menu_data, index=menu_items)
    
    print("--- Initial Menu Data ---")
    print(df)
    print("\n")


    # 2. Retrieving value via at (Label Specification)
    print("=== Retrieval via at (Label) ===")
    
    # Get Price of Mocha
    # Format: df.at[row_label, column_name]
    mocha_price = df.at["Mocha", "Price"]
    
    print(f"Price of Mocha: {mocha_price} Yen")


    # 3. Retrieving value via iat (Position Specification)
    print("\n=== Retrieval via iat (Position) ===")
    
    # Get Calories of Cappuccino
    # Cappuccino is 4th from top -> Row number 3
    # Calories is 2nd from left -> Column number 1
    # Format: df.iat[row_number, col_number]
    cappuccino_cal = df.iat[3, 1]
    
    print(f"Calories of Cappuccino: {cappuccino_cal} kcal")


    # 4. Updating Data
    print("\n=== Updating Values ===")
    
    # Update Price of Americano
    print(f"Before Update Americano Price: {df.at['Americano', 'Price']}")
    
    # Assign value using at (Price revision)
    df.at["Americano", "Price"] = 400
    
    print(f"After Update Americano Price: {df.at['Americano', 'Price']}")
    
    print("\n--- DataFrame After Update ---")
    print(df)

if __name__ == "__main__":
    access_scalar_values()

Execution Result

--- Initial Menu Data ---
            Price  Calories
Espresso      350        10
Latte         450       150
Mocha         500       280
Cappuccino    450       120
Americano     380        15


=== Retrieval via at (Label) ===
Price of Mocha: 500 Yen

=== Retrieval via iat (Position) ===
Calories of Cappuccino: 120 kcal

=== Updating Values ===
Before Update Americano Price: 380
After Update Americano Price: 400

--- DataFrame After Update ---
            Price  Calories
Espresso      350        10
Latte         450       150
Mocha         500       280
Cappuccino    450       120
Americano     400        15

Explanation: Differences from loc/iloc

Usually, loc and iloc are used for data access, but these are multi-functional accessors capable of handling “range selection” and retrieving “multiple rows/columns.”

In contrast, at and iat are specialized for accessing “single values (scalar values).” Therefore, when reading or writing individual values frequently within a loop process, using at / iat can be expected to improve processing speed.

df.at[row, col]: Use when you want to identify the location by label (name).
df.iat[row_int, col_int]: Use when you want to identify the location by number (position).

よかったらシェアしてね！