Describe the enhancement requested
Currently, the body of a Record Batch (i.e. the actual arrow arrays) is serialized before it's put into the internal representation. However, the Arrow JSON test data format requires this data to be symbolically represented in JSON. Thus, we need to maintain a symbolic representation in the IR to consume and produce Arrow JSON.
Needed by #74 and #75.
Describe the enhancement requested
Currently, the body of a Record Batch (i.e. the actual arrow arrays) is serialized before it's put into the internal representation. However, the Arrow JSON test data format requires this data to be symbolically represented in JSON. Thus, we need to maintain a symbolic representation in the IR to consume and produce Arrow JSON.
Needed by #74 and #75.