aboutsummaryrefslogtreecommitdiff
path: root/core
diff options
context:
space:
mode:
authorWenchen Fan <wenchen@databricks.com>2016-01-31 22:43:03 -0800
committerDavies Liu <davies.liu@gmail.com>2016-01-31 22:43:03 -0800
commitc1da4d421ab78772ffa52ad46e5bdfb4e5268f47 (patch)
treeda4be6e0c6142592d845daa5bc0928b7000c3769 /core
parent5a8b978fabb60aa178274f86432c63680c8b351a (diff)
downloadspark-c1da4d421ab78772ffa52ad46e5bdfb4e5268f47.tar.gz
spark-c1da4d421ab78772ffa52ad46e5bdfb4e5268f47.tar.bz2
spark-c1da4d421ab78772ffa52ad46e5bdfb4e5268f47.zip
[SPARK-13093] [SQL] improve null check in nullSafeCodeGen for unary, binary and ternary expression
The current implementation is sub-optimal: * If an expression is always nullable, e.g. `Unhex`, we can still remove null check for children if they are not nullable. * If an expression has some non-nullable children, we can still remove null check for these children and keep null check for others. This PR improves this by making the null check elimination more fine-grained. Author: Wenchen Fan <wenchen@databricks.com> Closes #10987 from cloud-fan/null-check.
Diffstat (limited to 'core')
0 files changed, 0 insertions, 0 deletions