CODEWITHSUNDEEP
pandas
BossRoyce
BossRoyce

Reputation: 211

String Search Function

So this works:

idocs[idocs['Message'].str.contains('No sold-to found')]

But this doesn't:

def msg_logic(msg):
    if msg.str.contains('No sold-to found'):
       return 'No Sold-to'
    else:
       return 'Other'

idocs['Category'] = idocs.apply(lambda x: msg_logic(x['Message']),axis=1)
print(idocs['Category'].unique())    
idocs.head()

Upvotes: 1

Views: 37

Answers (3)

Rafael Ferreira
Rafael Ferreira

Reputation: 328

def msg_logic(msg):
    if 'No sold-to found' in msg:
       return 'No Sold-to'
    else:
       return 'Other'

idocs['Category'] = idocs.apply(lambda x: msg_logic(x['Message']),axis=1)
print(idocs['Category'].unique())    
idocs.head()

Your code does not run because msg in the function is a string no a dataframe.

or 

idocs['Category'] = idocs['Message'].apply(lambda x: "No Sold-to" if "No sold-to found" in x else "Other")

Upvotes: 1

rafaelc
rafaelc

Reputation: 59274

I recommend np.where

idocs['Category'] = (np.where(idocs['Message'].str.contains('No sold-to found'),
                             'No Sold-to',
                             'Other'))

For multiple conditions, use np.select

Upvotes: 2

CypherX
CypherX

Reputation: 7353

Try this. 

idocs['Category'] = idocs['Message'].str.contains('No sold-to found')
idocs['Category'] = idocs['Category'].apply(lambda x: 'No Sold-to' if x else 'Other')

Upvotes: 0

Related Questions