How to add map column in spark based on other column?

问题

I have this table:

|Name|Val|
|----|---|
|Bob |1  |
|Marl|3  |

And I want to transform it to a map with single element like this:

|Name|Val|MapVal|
|----|---|------|
|Bob |1  |(0->1)|
|Marl|3  |(0->3)|

Any idea how to do it in scala? I couldn't find any way to build a map in withColumn statement...

回答1:

Found it - Just need to include the implicit sql:

import org.apache.spark.sql.functions._

And then use the map function: df.withColumn("MapVal", map(lit(0), col("Val")))

来源：https://stackoverflow.com/questions/42466708/how-to-add-map-column-in-spark-based-on-other-column