yisun123456: df.approxQuantile(col='count',probabilities=[0.1,0.3,0.5,0.7,0.9],relativeError=0.00001)

枫叶-哈哈: 写着很详细，厉害

yisun123456: wind = Window.partitionBy(['name','addr']) med = F.expr('percentile_approx(len, array(0.25, 0.5, 0.75, 0.95))') #df.withColumn('med_val', med.over(wind)).show() spark.read.text("/user/data/my_name/rec/seq_outputs/{}".format(cur_date))\ .withColumn('len',F.split(col('value'),';')[4])\ .withColumn('len',col('len').cast(IntegerType()))\ .withColumn('med_val', med.over(wind)) \ .withColumn('rn',F.expr("row_number() over(partition by name.addr order by time)"))\ .where('rn==1')\ .orderBy([F.col('name'),F.col('addr')])\ .show()

sanmu400: 代码运行错误

dxyzzzzz: db.gpsInfo_c.find({"upload_time":{$type:9}}).forEach(function(x){x.upload_time=new Date(x.upload_time); db.gpsInfo_copy.save(x)}) > [Error] Invalid key '_id': update only works with $ operators and pipelines at line 3, column 3 > 时间: 0.006s 直接报这个错是什么情况啊大佬？