更改DataFrame大小时来自casefold（）的KeyError

Question

我正在使用一个项目的快速约会数据集，对其应用了不同的分类器算法。注意到匹配数（1380）远低于“不匹配”数（6998）后，我想看看分类器在数字相等时的性能，因此尝试调整所述数据集的大小。问题是调整大小后，.casefold()函数（在最后一个for循环中找到）返回KeyError：1666（这是最后一个“不匹配”的索引，然后才开始从DataFrame中删除项目），坦率地说，我不明白为什么。

speed = pd.read_excel('speeddatingdataset.xls')

# Resizing the data set

zeros = 0
for i in range(len(speed)):
    if(speed.at[i,'match'] == 0 and zeros <1381):
        zeros+=1
    if(zeros >= 1381 and speed.at[i,'match'] == 0):
        speed = pd.DataFrame.drop(speed, labels = i, axis = 0)

# Preprocessing of the Data Set stage

match = speed['match'] # save the class values separately before removing from dataframe

speed = pd.DataFrame.drop(speed,columns=['has_null','wave','match','decision','decision_o']) # dropped the has_null and wave columns, as they don't add information

def replace_race(i,fName):
    if(speed.at[i,fName] == "?"):
        speed.at[i,fName] = 0
    elif(speed.at[i,fName] == "European/Caucasian-American"):
        speed.at[i,fName] = 1
    elif(speed.at[i,fName] == "'Black/African American'"):
        speed.at[i,fName] = 2
    elif(speed.at[i,fName] == "'Latino/Hispanic American'"):
        speed.at[i,fName] = 3
    elif(speed.at[i,fName] == "'Asian/Pacific Islander/Asian-American'"):
        speed.at[i,fName] = 4
    elif(speed.at[i,fName] == "Other"):
        speed.at[i,fName] = 5

def replace_field():
    proffesions=list()
    for i in range(0,len(speed)):
        if(len(proffesions) == 0):
            proffesions.append(speed.at[i,'field'])
            speed.at[i,'field'] = proffesions.index(speed.at[i,'field'])
        elif(proffesions.count(speed.at[i,'field']) > 0):
            speed.at[i,'field'] = proffesions.index(speed.at[i,'field'])
        elif(proffesions.count(speed.at[i,'field']) == 0):
            proffesions.append(speed.at[i,'field'])
            speed.at[i,'field'] = proffesions.index(speed.at[i,'field'])

for i in range(0,len(speed)):
    speed.at[i,'field'] = speed.at[i,'field'].casefold()
    if (speed.at[i,'field'].startswith('"') or speed.at[i,'field'].startswith('\'')):
        speed.at[i,'field'] = speed.at[i,'field'][1:]
    if (speed.at[i,'field'].endswith('"') or speed.at[i,'field'].endswith('\'')):
        speed.at[i,'field'] = speed.at[i,'field'][:(len(speed.at[i,'field'])-1)]
    if (speed.at[i,'gender'] == '?') :
        speed.at[i,'gender'] = 0
    elif(speed.at[i,'gender'] == "female"):
        speed.at[i,'gender'] = 1
    elif(speed.at[i,'gender'] == "male"):
        speed.at[i,'gender'] = 2
    replace_race(i,'race')
    replace_race(i,'race_o')
replace_field()