如何在特定条件下将某个 function 应用于 DataFrame 列值

Question

I have the df below:

ID     Value
 1        45
 2      -111

df = pd.DataFrame(columns=['ID', 'Value'], data=[['1', '45'], ['2', '-111'])

How can I apply the np.radians() function to the Value column if and only if the current value does not equal -111?

Final output would be:

ID     Value
 1  0.785398
 2      -111

I am trying to do something like

df['Value'] = df.apply(lambda row: '-111' if row['Value'] == '-111' else np.radians(row['Value'].astype(float)), axis=1)

but am running into isssues

Answer 1

You can use .where() as follows:

df['Value'] = df['Value'].where(df['Value'] == -111, np.radians(df['Value']))

.where() keeps the values of the series if the test condition is true and replace by the values of the second parameter if the test condition is false. Thus, it replace the values only when the value not equals -111 here.

Result:

print(df)


   ID       Value
0   1    0.785398
1   2 -111.000000

Answer 2

You can use boolean-indexing:

# convert Value column to float
df["Value"] = df["Value"].astype(float)

# create a mask
mask = df.Value != -111

# apply np.radians using the mask
df.loc[mask, "Value"] = df.loc[mask, "Value"].apply(np.radians)
print(df)

Prints:

  ID       Value
0  1    0.785398
1  2 -111.000000

Answer 3

np.where option:

df.Value = np.where(df.Value.ne(-111), radians(df.Value) , df.Value)

Full code:

from numpy import radians
df.Value = df.Value.astype(int)
df.Value = np.where(df.Value.ne(-111), radians(df.Value) , df.Value)