Tag: 値

Extracting a single specific value from a pandas DataFrame in Python is surprisingly tricky.
DataFrames are designed to be handled as DataFrames, so pulling out a row, slicing, or appending data is very easy.

But when you try to extract a value, for some reason it does not come out as a value—it stays in DataFrame form, and various other problems come up.

Here I describe how to extract the value in a DataFrame that satisfies a particular condition.
There are probably better ways to do this, but this is the best I can manage at my current level.

First, suppose we have the DataFrame below and want to extract the value 20 outlined in red.

The catch is that with a simple DataFrame like the one above you can see the index and pull the value out in one step, but with a large dataset it is very hard to check the index number and use it.

With a large dataset, it is more efficient to extract data using a key column (customer number, ID, etc.) as a clue.
So let’s try extracting the value in the situation below.

In other words, suppose we want the value in column B for the row where column A is Aichi.
You cannot do this in one step; the value is extracted in two stages.

First, create a DataFrame like the one above.

import pandas as pd
df = pd.DataFrame({ 'A' : ["Tokyo", "Aichi", "Osaka"],
                    'B' : [10, 20, 30],
                    'C' : [100, 200, 300]})
df

Next, extract the row where column A is Aichi.

a = df[df["A"] == "Aichi"]

print(a)

This lets us pull out the row containing Aichi.
Finally, we extract the value in column B of this row.

b = a.at[a.index[0], "B"]

print(b)

Doing this, you can extract the value.
It is roundabout, but this is the method I currently use.

If you know of another approach, I would be glad to hear about it in the comments.

import pandas as pd
df = pd.DataFrame({ 'A' : ["Tokyo", "Aichi", "Osaka"],
                    'B' : [10, 20, 30],
                    'C' : [100, 200, 300]})

a = df[df["A"] == "Aichi"]
b = a.at[a.index[0], "B"]
print(b)

Here is a summary of the code above.