Error: Pandas Hashtable Keyerror

Error: pandas hashtable keyerror

I think first is best investigate, what are real columns names, if convert to list better are seen some whitespaces or similar:

print (reviews_new.columns.tolist())

---

I think there can be 2 problems (obviously):

1.whitespaces in columns names (maybe in data also)

Solutions are strip whitespaces in column names:

reviews_new.columns = reviews_new.columns.str.strip()

Or add parameter skipinitialspace to read_csv:

reviews_new = pd.read_csv("D:\\aviva.csv", skipinitialspace=True)

2.different separator as default ,

Solution is add parameter sep:

#sep is ;
reviews_new = pd.read_csv("D:\\aviva.csv", sep=';')
#sep is whitespace
reviews_new = pd.read_csv("D:\\aviva.csv", sep='\s+')
reviews_new = pd.read_csv("D:\\aviva.csv", delim_whitespace=True)

EDIT:

You get whitespace in column name, so need 1.solutions:

print (reviews_new.columns.tolist())
['Name', ' Date', ' review'] 
          ^        ^

Pandas Hashtable KeyError

The following lines

X = df['MaterialDescription']
clean_review = description_to_words(X[3] )

give in python description_to_words(df['MaterialDescription'][3] )

You have to locate your index through:

clean_review = description_to_words(df.iloc[3]['MaterialDescription'] )

pandas._libs.hashtable.PyObjectHashTable.get_item KeyError: 0

I think this error comes from your [i] notation, which is trying to look for the DataFrame index value of 0, which doesn't exist. Try replacing every instance of [i] with .iloc[i].

Also, you should be able to replace the for loop with much more compact, readable, and less error-prone code, especially since you define emotion_map but use it only for output. Try changing the mapping from strings to integers with emotion_map = { 0:6, 1:3, 2:4, 3:5, 4:2, 5:1, 6:0}, then move it to just under filtered_csv = ..., and replace that for loop with

filtered_csv['expression'] = filtered_csv['expression'].replace(emotion_map)

---

Related Topics