| This is an epic Jira to support GeoSpatial datatypes natively in Hive. This will cater to the applications which queries on large volumes of spatial data. This support will be added in a phased manner. To start with, we are planning to make use of framework developed by ESRI [https://github.com/Esri/spatial-framework-for-hadoop |https://github.com/Esri/spatial-framework-for-hadoop).] on the Apache 2.0 license. This project is not very active and there is no release done to maven central. So its not easy to get the jars downloaded directly using pom dependency. Also the UDFs are based on older version of Hive. So we have decided to make a copy of this repo and maintained it inside Hive. This will make it easier to do any improvement and manage dependencies. As of now, the data loading is done only on a binary data type. We need to enhance this to make it more user friendly. In the next phase, a native Geometry/Geography datatype will be supported. User can directly create a geometry type and operate on it. Apart from these we can start adding support for different indices like quad tree, R-tree, ORC/Parquet/Iceberg support etc. |
New Feature
Major
