编写Python自定义函数判断入参时触发pandas DataFrame真值歧义ValueError
报错原因
你用if df == None:的判断思路没有错,但写法存在硬伤,是这次报错的核心来源:
当df参数传入的是pandas DataFrame对象时,df == None会触发DataFrame重写的相等判断逻辑,逐元素比对DataFrame每个单元格的值和None的匹配关系,最终生成一个和原DataFrame维度完全一致的布尔矩阵。此时if语句需要把这个矩阵转换为单个布尔值做真假判断,pandas无法定义多元素矩阵的整体真值,就会抛出“真值存在歧义”的报错。
判断对象是否为None的标准逻辑,是直接比对对象的内存地址,而非做值层面的相等判断。
解决方案
- 把函数内所有对None的判断,从
变量 == None替换为变量 is None:- 替换
if df == None:为if df is None: - 替换
if df2 == None:为if df2 is None:
这种写法不会触发任何DataFrame内部的运算逻辑,会直接精准判断变量是否指向None对象,从根源上规避报错。
- 替换
- 修正代码中的笔误:在df2为DataFrame对象的分支里,打印前10条记录的那行变量名错写为
ddf_in2_in,需要改回df_in2,否则后续调用传入DataFrame类型的df2参数时,会触发变量未定义的NameError。
修正后可直接运行的代码示例
import pandas as pd DATAPATH = "path/" df_orig = pd.read_csv(DATAPATH + "file.csv" , dtype=str, sep = "|") def test(df = None, df2 = None): if df is None: pass elif not isinstance(df, pd.DataFrame): df_in = pd.read_csv(DATAPATH + df + ".csv", dtype=str, sep = "|") print(f'============ Number of obs. in {df} ============\n{df_in.shape[0]:,}\n') print(f'============ First 10 Records of {df}.csv ============\n{df_in.head(10)}\n\n\n') print(f'============ Last 10 Records of {df}.csv ============\n{df_in.tail(10)}\n\n\n') else: df_in = df print(f'============ Number of obs. ============\n{df_in.shape[0]:,}\n') print(f'============ First 10 Records ============\n{df_in.head(10)}\n\n\n') print(f'============ Last 10 Records ============\n{df_in.tail(10)}\n\n\n') if df2 is None: pass elif not isinstance(df2 , pd.DataFrame): df_in2 = pd.read_csv(DATAPATH + df2 + ".csv", dtype=str, sep = "|") print(f'============ Number of obs. in {df2} ============\n{df_in2.shape[0]:,}\n') print(f'============ First 10 Records of {df2}.csv ============\n{df_in2.head(10)}\n\n\n') print(f'============ Last 10 Records of {df2}.csv ============\n{df_in2.tail(10)}\n\n\n') else: df_in2 = df2 print(f'============ Number of obs. ============\n{df_in2.shape[0]:,}\n') print(f'============ First 10 Records ============\n{df_in2.head(10)}\n\n\n') print(f'============ Last 10 Records ============\n{df_in2.tail(10)}\n\n\n') test(df = df_orig, df2 = None)
内容的提问来源于stack exchange,提问作者Peter Chen
相关产品推荐
相关产品推荐

