Creating a new column based on the values of other columns

Question

I wanted to create a "High Value Indicator" column, which says "Y" or "N" based on two different value columns. I want the new column to have a "Y" when Value_1 is > 1,000 or Value_2 > 15,000. Bellow is the table, the desired output would include the indicator column based on the or condition about.

ID   Value_1     Value_2 
1    100         2500
2    250         6250
3    625         15625
4    1500        37500
5    3750        93750

Answer 1

Try using .loc and .fillna

df.loc[((df['Value_1'] > 1000) 
       |(df['Value_2'] > 15000)), 'High_Value_Ind'] = 'Y'

df['High_Value_Ind'] = df['High_Value_Ind'].fillna('N')

Answer 2

Use numpy.where with chained conditions by | for or:

df['High Value Indicator'] = np.where((df.Value_1 > 1000) | (df.Value_2 > 15000), 'Y', 'N')

Or map by dictionary:

df['High Value Indicator'] = ((df.Value_1 > 1000) | (df.Value_2 > 15000))
                                 .map({True:'Y', False:'N'})

print (df)
   ID  Value_1  Value_2 High Value Indicator
0   1      100     2500                    N
1   2      250     6250                    N
2   3      625    15625                    Y
3   4     1500    37500                    Y
4   5     3750    93750                    Y

Timings:

df = pd.concat([df] * 10000, ignore_index=True)

---

In [76]: %timeit df['High Value Indicator1'] = np.where((df.Value_1 > 1000) | (df.Value_2 > 15000), 'Y', 'N')
100 loops, best of 3: 4.03 ms per loop

In [77]: %timeit df['High Value Indicator2'] = ((df.Value_1 > 1000) | (df.Value_2 > 15000)).map({True:'Y', False:'N'})
100 loops, best of 3: 4.82 ms per loop

In [78]: %%timeit
    ...: df.loc[((df['Value_1'] > 1000) 
    ...:        |(df['Value_2'] > 15000)), 'High_Value_Ind3'] = 'Y'
    ...: 
    ...: df['High_Value_Ind3'] = df['High_Value_Ind3'].fillna('N')
    ...: 
100 loops, best of 3: 5.28 ms per loop


In [79]: %timeit df['High Value Indicator'] = (df.apply(lambda x: 'Y' if (x.Value_1>1000 or x.Value_2>15000) else 'N', axis=1))
1 loop, best of 3: 1.72 s per loop