Github user WeichenXu123 commented on the issue: https://github.com/apache/spark/pull/19439 @hhbyyh Thanks for your comments! > Another option is that to support all bytes[], short[], int[], float[] and double[] as data storage type candidates, and switch among them according to CvType Do you mean add more columns in the `image` struct ? Each type a column ? And any time there will only be one column not null ? It make schema redundant. > Not sure about the reason to include "origin" info into the image data. What about making "origin" as an optional column, i.e, `readImages` add a param to specify whether to fill "origin" column ? > Does the wildcard matching still work if recursive == false ? I think it still works. This is implemented in hadoop code.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org