问题
If you want to create a pipe with more than 22 fields from a smaller one in Scalding you are limited by Scala tuples, which cannot have more than 22 items.
Is there a way to use collections instead of tuples? I imagine something like in the following example, which sadly doesn't work:
input.read.mapTo('line -> aLotOfFields) { line: String =>
(1 to 24).map(_.toString)
}.write(output)
回答1:
actually you can. It's in FAQ - https://github.com/twitter/scalding/wiki/Frequently-asked-questions#what-if-i-have-more-than-22-fields-in-my-data-set
val toFields = (1 to 24).map(f => Symbol("field_" + f)).toList
input
.read
.mapTo('line -> toFields) { line: String =>
new Tuple((1 to 24).map(_.toString).map(_.asInstanceOf[AnyRef]): _*)
}
the last map(_.asInstanceOf[AnyRef]) looks ugly so if you find better solution let me know please.
回答2:
Wrap your tuples into case classes. It will also make your code more readable and type safe than using tuples and collections respectively.
来源:https://stackoverflow.com/questions/19593137/can-i-output-a-collection-instead-of-a-tuple-in-scalding-map-method