...
Whether enable loading UDFs from metastore on demand; this is mostly relevant for HS2 and was the default behavior before Hive 1.2.
mapreduce.input.fileinputformat.split.maxsize
Default Value: empty
For splittable data this changes the portion of the data that each mapper is assigned. By default each mapper is assigned based off of the block sizes of the source files. Entering a value larger than the block size will will decrease the number of splits which creates less mappers. Entering a value smaller than the block size will increase the number of splits which creates more mappers.
SerDes and I/O
SerDes
hive.script.serde
...
For WebHCat configuration, see Configuration Variables in the WebHCat manual.