scala filter操作的记录（持续增加），创建空表

1.filter过滤指定字符串的值

val a = lines.filter(x=>x.contains("python")) //选出包含“python”的行

2.Dataframe filter sql语法正则匹配

val a = df.filter("columnName rlike '正则'")

3.Dataframe filter 过滤列表内值

val dateList = List（1,2,3,4,5）
val a = df.filter('dt.isInCollection(dateList))

4.sql过滤数组(选出数组内的指定渠道)

df.filter("channel in (2,121,53,129)

5.创建空表+空列

spark.emptyDataFrame.toDF().
      selectExpr("'0' as channel", "'0' as time", "'0' as ip", "'0' as long", "'0' as uin","'0' as version", "'0' as country", "'0' as language", "'0' as id", "'0' as platform")