CODEWITHSUNDEEP
pythonpandasdataframe
Jessica Chambers
Jessica Chambers

Reputation: 1316

Concatenating a row in a pandas Dataframe

This seems like it should be so much simpler yet here I am.

I'm trying to add a row to a data frame (2 data frames to be exact) from another data frame, but I get the following error:

TypeError: cannot concatenate object of type "<class 'numpy.float64'>"; only pd.Series, pd.DataFrame, and pd.Panel (deprecated) objs are valid

My code

for i in range(0,len(k_means_labels_unique)):
    X = pd.DataFrame(columns=['first occurrence of \'AB\'','similarity to \'AB\''])
    y = pd.DataFrame(columns=['Class'])
    for row in result.iterrows():
        data=row[1]
        if data['cluster ID'] == i:
            X = pd.concat([X,data['first occurrence of \'AB\''],data['similarity to \'AB\'']])
            y = pd.concat([y,data['Class']])

Do I have to transform data['first occurrence of \'AB\''],data['similarity to \'AB\''] into another data frame? This seems horribly inefficient

EDIT: I tried y = pd.concat([y,pd.Series(data['Class'])]) but that appended the data as a new column, example for y:

columns

Upvotes: 4

Views: 22899

Answers (1)

CoMartel
CoMartel

Reputation: 3591

You need to first convert to dataframe :

X = pd.concat([X,pd.DataFrame([[data['first occurrence of \'AB\''],data['similarity to \'AB\'']]],columns=['first occurrence of \'AB\'','similarity to \'AB\''])], ignore_index=True)
y = pd.concat([y,pd.DataFrame([data['Class']], columns=['Class'])], ignore_index=True)

EDIT : add ignore_index=True

Upvotes: 2

Related Questions