The nested fields in doc such as {key1: value1, key2: {key21: value21, key22: value22}} that are derived as string since they could be of different data types, get converted to {key1: value1, key2: {key21 = value21, key22 = value22}} when written out as parquet. The hive json parser functions such as "get_json_object" is not able to parse this and needs to be in standard json format.
Added the boolean config "ConvertNestedDocsToNativeJsonFormat" with default of false to enable this.
The nested fields in doc such as {key1: value1, key2: {key21: value21, key22: value22}} that are derived as string since they could be of different data types, get converted to {key1: value1, key2: {key21 = value21, key22 = value22}} when written out as parquet. The hive json parser functions such as "get_json_object" is not able to parse this and needs to be in standard json format.
Added the boolean config "ConvertNestedDocsToNativeJsonFormat" with default of false to enable this.