Pandas代码报错：timedelta64[ns]与int无效比较的解决方法

问题描述 #

运行以下Python Pandas代码时触发TypeError：

for file in files:
        df=pd.read_csv(file)
        df["time_stamp"]=pd.to_datetime(df["time_stamp"], infer_datetime_format=True)
        df= df.resample('1Min', on='time_stamp').first().dropna()
        df["time_diff"]=df["time_stamp"].diff() 
        df=df[1:]
        df["time_diff_h"]=df["time_diff"].apply(lambda x: x.total_seconds()/3600)
        
        
        df=df.reset_index(drop=True)
        #index_list=df[df["time_diff_h"]>time_gap].index
        index_list = df[df["time_diff_h"]>24].index
        
        if len(index_list)==0:
            continue
        df_output=pd.DataFrame()
        for i in index_list:
            soc_diff=df["ibs3_soc__0x7a3 (%)"].iloc[i]-df["ibs3_soc__0x7a3 (%)"].iloc[i-1]
            initial_soc=df["ibs3_soc__0x7a3 (%)"].iloc[i-1]
            final_soc=df["ibs3_soc__0x7a3 (%)"].iloc[i]
            timeDiff= round((df["time_diff_h"].iloc[i]),2)
            initial_time=df["time_stamp"].iloc[i-1]
            final_time=df["time_stamp"].iloc[i]
            power=(df["ibs3_q_released__0x7a3 (ah)"].iloc[i]-df["ibs3_q_released__0x7a3 (ah)"].iloc[i-1])/timeDiff
           
            if power<0.125:
                continue
            else:
                features={"File": file, "SOC_Difference (%)": soc_diff, "Initial_SOC (%)": initial_soc, "Final_SOC (%)": final_soc, "Time_Diff (h)": timeDiff, "Initial_time": initial_time, "Final_time": final_time, "Power_released (Ah)": power}
            
            df_output=df_output.append(pd.Series(features), ignore_index=True)

报错信息：

raise TypeError(f"Invalid comparison between dtype={left.dtype} and {typ}")

TypeError: Invalid comparison between dtype=timedelta64[ns] and int

已通过代码排查确认仅time_diff列为timedelta64[ns]类型，且未直接对该列执行数值比较，无法定位报错原因，寻求解决方案。

解决方法 #

1. 替换apply为矢量化时间差转换 #

apply属于逐元素操作，容易因个别异常值导致列类型混乱。改用Pandas内置的dt访问器做矢量化转换，确保输出为数值类型：

# 替换原time_diff_h生成代码
df["time_diff_h"] = df["time_diff"].dt.total_seconds() / 3600

2. 修正resample后的索引处理 #

resample操作后time_stamp会成为索引，后续直接调用df["time_stamp"].diff()可能触发隐式类型转换异常。调整resample代码，显式重置索引：

df = df.resample('1Min', on='time_stamp').first().dropna().reset_index()

3. 强制比较两边类型一致 #

在生成index_list时，将整数24显式转为浮点型，避免类型匹配问题：

index_list = df[df["time_diff_h"]>24.0].index

4. 临时排查：打印关键列类型 #

在报错代码前添加打印语句，确认time_diff_h的实际类型和样本值，快速定位异常：

print("time_diff_h dtype:", df['time_diff_h'].dtype)
print("time_diff_h sample:", df['time_diff_h'].head())

内容的提问来源于stack exchange，提问作者AC24