Question

我在pyspark脚本行下面，

df_output = df.select("*",$checkcol) 
df_output.show()

通过对变量进行硬编码可以很好地工作，但是参数化后会抛出错误，

pyspark.sql.utils.AnalysisException：'无法解析\'`**，F .....

其中checkcol是一个变量，其值如下所示，

checkcol-

F.when(F.col("colA")=='null',"Yes").otherwise(date_validation_udf("colA")).alias("colA_DateCheck"),
F.when(F.col("colB")=='null',"Yes").otherwise(date_validation_udf("colB")).alias("colB_DateCheck"),F.when(F.col("colC")=='null',"Yes").otherwise(date_validation_udf("colC")).alias("colC_DateCheck"),
F.when(F.col("colD")=='null',"Yes").otherwise(num_check_udf("colD")).alias("colD_NumCheck"),F.when(F.col("colE")=='null',"Yes").otherwise(num_check_udf("colE")).alias("colE_NumCheck"),
F.when(F.col("colF")=='null',"Yes").otherwise(num_check_udf("colF")).alias("colF_NumCheck"),F.when(F.col("colG")=='null',"Yes").otherwise(num_check_udf("colG")).alias("colG_NumCheck")

Answer 1

尝试一下：

import pyspark.sql.functions as F

df_output = df.withColumn("colA",
                          F.when(F.col("colA")=='null',"Yes").otherwise(date_validation_udf("colA")).alias("colA_DateCheck"))
              .withColumn("colB",
                          F.when(F.col("colB")=='null',"Yes").otherwise(date_validation_udf("colB")).alias("colB_DateCheck"),F.when(F.col("colC")=='null',"Yes").otherwise(date_validation_udf("colC")).alias("colC_DateCheck"),)
              ...

df_output.show()

编辑：

要将这些语句作为一个变量传递，请尝试以下操作：

checkcol = (F.when(F.col("colA") == 'null', "Yes").otherwise(date_validation_udf("colA")).alias("colA_DateCheck"),
            F.when(F.col("colB") == 'null', "Yes").otherwise(date_validation_udf("colB")).alias("colB_DateCheck"),
            F.when(F.col("colC") == 'null', "Yes").otherwise(date_validation_udf("colC")).alias("colC_DateCheck"),
            F.when(F.col("colD") == 'null', "Yes").otherwise(num_check_udf("colD")).alias("colD_NumCheck"),
            F.when(F.col("colE") == 'null', "Yes").otherwise(num_check_udf("colE")).alias("colE_NumCheck"),
            F.when(F.col("colF") == 'null', "Yes").otherwise(num_check_udf("colF")).alias("colF_NumCheck"),
            F.when(F.col("colG") == 'null', "Yes").otherwise(num_check_udf("colG")).alias("colG_NumCheck"))


df_output =  df.select(
          '*', 
           *checkcol
)

AnalysisException，pyspark无法解析数据框查询中的变量

1 个答案: