CODEWITHSUNDEEP
pythonpandasnumpywhere-clause
DeAnna
DeAnna

Reputation: 402

Need a conditional statement to fill a column based off a string

I need to fill cells in a column based off if another column contains a certain string.

I need to fill column B based off what's in C. Like if C contains 'hello;', then fill the corresponding cell in B with 'greet'. Then if C contains 'bye;', fill the corresponding cells in B with 'farewell'. 

df1
   A    B       C        D
0  w       hello; Jon    q
1  x       bye; Jon      r
2  y       hello; Jack   s
3  z       bye; Jack     t

df1['B'] = np.where(df1['C'].str.contains('hello;'), 'greet', '')
df1['B'] = np.where(df1['C'].str.contains('bye;'), 'farewell', '')

This works; however, the next line of code overwrites the 'greet' from the first line. So I'm not sure how to combine the conditionals so they don't overwrite each other. What I want the end result to be is 

df1
   A    B          C          D
0  w  greet     hello; Jon    q
1  x  farewell  bye; Jon      r
2  y  greet     hello; Jack   s
3  z  farewell  bye; Jack     t

Upvotes: 1

Views: 763

Answers (2)

dijksterhuis
dijksterhuis

Reputation: 1271

If you’re only going to be dealing with a binary choice and all the values exist in the column, as per the example, then this should be fine:

 df1['B'] = np.where(df1['C'].str.contains('bye;'), 'farewell', 'greet')

From the numpy docs:

numpy.where(condition[, x, y])

Return elements chosen from x or y depending on condition.

If the condition is satisfied, it will return x, else it will fill with y.

However, np.select docs will be the one you want if you have more than one condition:

conditions = [
    df['C'].str.contains('hello;'),
    df['C'].str.contains('bye;')
]

np.select(conditions, ['greet', 'farewell'])


array(['greet', 'farewell', 'greet', 'farewell'], dtype='<U11')

Upvotes: 2

tawab_shakeel
tawab_shakeel

Reputation: 3739

try using np.select

m1= df['C'].str.contains('hello;')
m2= df['C'].str.contains('bye;')

df['B'] = np.select(condlist=[m1 , m2], 
                         choicelist=['greet','farewell'])

Upvotes: 1

Related Questions