srowen commented on a change in pull request #24362: [SPARK-27454][ML][SQL]
Spark image datasource fail when encounter some illegal images
URL: https://github.com/apache/spark/pull/24362#discussion_r275466292
##########
File path: mllib/src/main/scala/org/apache/spark/ml/image/ImageSchema.scala
##########
@@ -133,7 +133,13 @@ object ImageSchema {
*/
private[spark] def decode(origin: String, bytes: Array[Byte]): Option[Row] =
{
- val img = ImageIO.read(new ByteArrayInputStream(bytes))
+ val img = try {
+ ImageIO.read(new ByteArrayInputStream(bytes))
+ } catch {
+ // Catch runtime exception because `ImageIO` may throw unexcepted
`RuntimeException`.
+ // But do not catch the declared `IOException` (regarded as FileSystem
failure)
+ case _: RuntimeException => null
Review comment:
I suppose in other cases we have a new column for 'bad records' but not sure
if that's overkill or what. I'm OK with just making this change here and doing
more later as it already ignores some images that ImageIO doesn't read
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
