Our schema is nested with top level having 5 struct types. When we try to query these structs we get the following back
*ORC does not support type conversion from file type string (1) to reader type array<string> (1)* Walking through hive in a debugger I see that schema evolution sees the correct file type (which matches schema), but the reader type looks like the following *struct<_col0:array<string>>* minified file type *struct<id:string,source:string,timestamp:bigint,data:struct<...>,output:struct<browser:string,provider:array<string>,...>,...>* So what I see is that schema evolution is trying to compare id to output.provider (or really _col0) which will fail since they don't match. query is select output.provider from table; hive version is 2.1.1 Any idea whats going on?
