python - Pandas - CSV 到 Dataframe(使用 Base64 编码的列)

Question

以下代码用于使用 Pandas 将防火墙日志从 csv 提取到数据帧中。

df = pd.read_csv('/Users/alistairgillespie/Documents/Projects/COMP5310/Akamai Data/FINAL/data.csv', dtype = {"_time": str, "city": str,"country": str,"lat": str,"long": str,"region": str,"UA": str,"bytes": str,"cliIP": str,"reqHost": str, "reqMethod": str, "reqPath": str,"reqPort": str,"respCT": str,"respLen": str,"status": str,"referer": str,"date": str,"conn": str,"denyData": str,"denyRules": str,"policy": str,"ruleSet": str,"warnRules": str,"warnData": str,"warnSlrs": str,"warnTags": str})

* 请原谅长行列

进入数据帧后，我想迭代每一行并使用 unquote 和 base64decode 函数调用解码“denyData”列字段(如果不是 NaN)。我正在尝试使用以下代码来执行此操作:

for i, row in df.iterrows():
    print(pd.notnull(row))
    temp = parse.unquote(row['denyData'])
    new = base64.b64decode(temp)
    df2.loc[i, 'denyData'] = new

产生以下错误:

TypeError: argument of type 'float' is not iterable

将 csv 字节列处理为 Pandas 数据帧的正确方法是什么？这是清理此类数据的正确方法吗？数据示例如下。

Answer 1

你可以尝试if-else，因为错误显然意味着无法处理NaN:

for i, row in df.iterrows():
    print(pd.notnull(row))
    if pd.notnull(row):
        df.loc[i, 'denyData'] = base64.b64decode(parse.unquote(row['denyData']))
    else:
        df.loc[i, 'denyData'] = np.nan