CODEWITHSUNDEEP
pythonpython-3.xpandas
Manu Mohan
Manu Mohan

Reputation: 177

Dataframe.str Fails as entire column is blank

I have data frame with many columns. While execution i got an error like 

Can only use .str accessor with string values, which use np.object_ dtype in pandas

I figured it out this is because one my column has only Nan values on that particular occasion and the .str failed on the code

df1['Merchant_old']=df1['Merchant_old'].str.lower().str.strip()

I don't how I should use isna() to run the above step only when the entire column is not blank.

Upvotes: 1

Views: 73

Answers (2)

hilberts_drinking_problem
hilberts_drinking_problem

Reputation: 11602

I assume that type of your df.col is np.float. You could get away with using str with df.col.astype(np.object_).str. This will keep np.NaNs in the final output.

Upvotes: 2

BENY
BENY

Reputation: 323226

You can create a mask here, with isnull and all

df=pd.DataFrame({'A':['A',np.nan,'B'],'B':[np.nan,np.nan,np.nan]})

s=df.isnull().all()
df.loc[:,s[~s].index]=df.loc[:,s[~s].index].apply(lambda x : x.str.lower().str.strip())
df
     A   B
0    a NaN
1  NaN NaN
2    b NaN

Upvotes: 1

Related Questions