How to avoid decoding to str: need a bytes-like object error in pandas?

Question

Here is my code :

data = pd.read_csv('asscsv2.csv', encoding = "ISO-8859-1", error_bad_lines=False);
data_text = data[['content']]
data_text['index'] = data_text.index
documents = data_text

It looks like

print(documents[:2])
                                              content  index
 0  Pretty extensive background in Egyptology and ...      0
 1  Have you guys checked the back end of the Sphi...      1

And I define a preprocess function by using gensim

stemmer = PorterStemmer()
def lemmatize_stemming(text):
    return stemmer.stem(WordNetLemmatizer().lemmatize(text, pos='v'))
def preprocess(text):
    result = []
    for token in gensim.utils.simple_preprocess(text):
        if token not in gensim.parsing.preprocessing.STOPWORDS and len(token) > 3:
            result.append(lemmatize_stemming(token))
    return result

And when I use this function:

processed_docs = documents['content'].map(preprocess)

It appears

TypeError: decoding to str: need a bytes-like object, float found

How to encode my csv file to byte-like object or how to avoid this kind of error?

Answer 1

Your data has NaNs(not a number).

You can either drop them first:

documents = documents.dropna(subset=['content'])

Or, you can fill all NaNs with an empty string, convert the column to string type and then map your string based function.

documents['content'].fillna('').astype(str).map(preprocess)

This is because your function preprocess has function calls that accept string only data type.

Edit:

How do I know that your data contains NaNs? Numpy nan are considered float values

>>> import numpy as np
>>> type(np.nan)
<class 'float'>