Details
-
Improvement
-
Status: Resolved
-
Trivial
-
Resolution: Won't Fix
-
0.10.0
-
None
-
None
Description
Currently, LazySimpleSerde cannot handle columns including newline character cause hadoop splits rows by newline character. If the overhead counting fields by full scan and merging partial lines is tolerable, multi-lined column can be reconstructed at runtime.
But with this method, multi-lined column should not be located at the last of the row. This is just a idea for HIVE-1898.