The script streaming_build.sh has been removed in 1.6.0, since Kylin
introduces the new Streaming design. The old(experimental and deprecated)
one has been removed. To learn how to use it, please refer to
http://kylin.apache.org/docs16/tutorial/cube_streaming.html
2016-12-02 19:30 GMT+08:00 汪胜 :
hongbin ma created KYLIN-2246:
-
Summary: redesign the way to decide layer cubing reducer count
Key: KYLIN-2246
URL: https://issues.apache.org/jira/browse/KYLIN-2246
Project: Kylin
Issue Type: Imp
Hello, I download kylin 1.6.0, and installed successful. Streaming cubing build
need script "streaming_build.sh" in ${KYLIN_HOME}/bin, but I didn't found this
file in kylin 1.6.0 source code. Besides, in kylin 1.6.0 source code, script
"kylin.sh" didn't handle the situation of "./kylin.sh stream
:) We have some similiar problem with non-IT
"Why do you want a result table of 1 Millon of rows ¿? "
I don't Know any solution.
The Q is: this is a problem/bug/issue of Kylin ? I think NO, the target of
this is protect to coprocesor (is mandatory)
Select * is not addecuate for Kylin
Impala,
I remember hongbin has further optimization on the limit push down after
1.5.3; @hongbin, can you confirm this?
2016-12-02 17:02 GMT+08:00 alaleiwang :
> i can see each regionserver scan over 2000+
>
> the query "select *" is called from tableau used by non-technical
> analyst,so we can'
i can see each regionserver scan over 2000+
the query "select *" is called from tableau used by non-technical
analyst,so we can't force them to bypass such clause or to do something
from tableau ...
how can "limit" work for "select *" without "where"?
--
View this message in context