You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
如何序列化所有几何?
我添加了自定义的几何图形,希望在spark程序中应用。由于目前不支持自定义几何的序列化,或者说我不知道如何使用FlatBuffers支持序列化我添加的class。导致目前每个spark程序的实例都需要重新构建一次rtree,这样效率很低。我尝试设置Property : spark.serializer=org.apache.spark.serializer.KryoSerializer 修改默认的序列化方式,但是没有作用,甚至很久都没有完成rtree的构建。有没有一种序列化方式,可以不受限制,任意添加自定义几何,且能够完成快速重建rtree?或者我应该怎样使FlatBuffers支持序列化我添加的class?
hi:
I added a custom geometric figure, hoping to apply it in the spark program. Since the serialization of custom geometries is currently not supported, or I don't know how to use FlatBuffers to support serialization of the classes I added. As a result, each instance of the spark program needs to rebuild the rtree, which is very inefficient. I tried to set Property: spark.serializer=org.apache.spark.serializer.KryoSerializer to modify the default serialization method, but it didn't work, and it didn't even complete the construction of rtree for a long time. Is there a serialization method that can add custom geometry without limitation, and can quickly rebuild the rtree?Or how should I make FlatBuffers support serialization of the class I added?
thank you!
The text was updated successfully, but these errors were encountered:
如何序列化所有几何?
我添加了自定义的几何图形,希望在spark程序中应用。由于目前不支持自定义几何的序列化,或者说我不知道如何使用FlatBuffers支持序列化我添加的class。导致目前每个spark程序的实例都需要重新构建一次rtree,这样效率很低。我尝试设置Property : spark.serializer=org.apache.spark.serializer.KryoSerializer 修改默认的序列化方式,但是没有作用,甚至很久都没有完成rtree的构建。有没有一种序列化方式,可以不受限制,任意添加自定义几何,且能够完成快速重建rtree?或者我应该怎样使FlatBuffers支持序列化我添加的class?
hi:
I added a custom geometric figure, hoping to apply it in the spark program. Since the serialization of custom geometries is currently not supported, or I don't know how to use FlatBuffers to support serialization of the classes I added. As a result, each instance of the spark program needs to rebuild the rtree, which is very inefficient. I tried to set Property: spark.serializer=org.apache.spark.serializer.KryoSerializer to modify the default serialization method, but it didn't work, and it didn't even complete the construction of rtree for a long time. Is there a serialization method that can add custom geometry without limitation, and can quickly rebuild the rtree?Or how should I make FlatBuffers support serialization of the class I added?
thank you!
The text was updated successfully, but these errors were encountered: