Github user dongjoon-hyun commented on the pull request: https://github.com/apache/spark/pull/11868#issuecomment-199982128 Oh, I see what is the point here now. @rxin , may I explain a little bit more? Mathematically, `NaN` equality is defined `false`. The followings are all **false** in Scala (and Java). * row_a(1).getDouble(0) == row_a(1).getDouble(0) * row_b(1).getFloat(0) == row_b(1).getFloat(0) * row_a(1).getDouble(0) == row_b(1).getFloat(0) * Double.NaN == Double.NaN * Float.NaN == Float.NaN * Double.NaN == Float.NaN However, Spark `Row` on master branch already returns **true** for the followings. * Row(Double.NaN) == Row(Double.NaN) * Row(Float.NaN) == Row(Float.NaN) As you guess easily, the followings are still **false**. * Row(Float.NaN).getFloat(0) == Row(Float.NaN).getFloat(0) * Row(Double.NaN).getDouble(0) == Row(Double.NaN).getDouble(0)
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org