我发现正确的语法埋在这个数据库后。
https://forums.databricks.com/questions/115/how-do-i-pass-parameters-to-my-sql-statements.html
您在查询前面添加一个小写的f，并在查询中用大括号括住变量名。

bkt = 1

prime = spark.sql(f"SELECT ((year(fdr_date))*100)+month(fdr_date) as fdr_year, count(*) as counts\
            FROM pwrcrv_tmp\
            where EXTR_CURR_NUM_CYC_DLQ={bkt}\
            and EXTR_ACCOUNT_TYPE in('PS','PT','PD','PC','HV','PA')\
            group by ((year(fdr_date))*100)+month(fdr_date)\
            order by ((year(fdr_date))*100)+month(fdr_date)")

prime.show(50)

这应该行得通

p_filename ='some value'
z='some value'

query = "INSERT into default.loginfordetails (filename,logdesc) values ('{}','{}') ".format(p_filename,z)
 
spark.sql(query)

由于您的查询使用多行查询，因此建议您采用良好的编码风格。为了更容易地使用""" """，并避免斜线一起作为一个很好的编码风格"\"，确实有很多问题或转换，以解析在python和pyspark一般或当使用nbconvert为笔记本到脚本和variables括号{}内的查询或使用.format(bkt)：
选件{}：

bkt=1;
prime = spark.sql(f"""SELECT ((year(fdr_date))*100)+month(fdr_date) as fdr_year, count(*) as counts
                FROM pwrcrv_tmp
                where EXTR_CURR_NUM_CYC_DLQ={bkt}
                and EXTR_ACCOUNT_TYPE in('PS','PT','PD','PC','HV','PA')
                group by ((year(fdr_date))*100)+month(fdr_date)
                order by ((year(fdr_date))*100)+month(fdr_date)""")

prime.show(50);

选件.format()：

bkt=1;
prime = spark.sql(f"""SELECT ((year(fdr_date))*100)+month(fdr_date) as fdr_year, count(*) as counts
                FROM pwrcrv_tmp
                where EXTR_CURR_NUM_CYC_DLQ={}
                and EXTR_ACCOUNT_TYPE in('PS','PT','PD','PC','HV','PA')
                group by ((year(fdr_date))*100)+month(fdr_date)
                order by ((year(fdr_date))*100)+month(fdr_date)""".format(bkt)

prime.show(50);