Best effort to find locations for input splits#2223
Conversation
|
@navis this is failing unit tests |
navis
force-pushed
the
besteffort-split-locations
branch
2 times, most recently
from
896259e to
3158684
Compare
|
@fjy fixed test fail |
There was a problem hiding this comment.
can you use org.apache.hadoop.mapreduce.lib.input.FileInputFormat?
There was a problem hiding this comment.
I've tried once but JobContext which needed for it is rather big object to create a dummy implementation. I prefer to use mapre.FileInputFormat for simplicity.
navis
force-pushed
the
besteffort-split-locations
branch
from
3158684 to
15759ca
Compare
There was a problem hiding this comment.
can you use log.error(..) instead?
navis
force-pushed
the
besteffort-split-locations
branch
from
15759ca to
2bf1f80
Compare
|
addressed comments |
There was a problem hiding this comment.
pls remove, or, i guess you need some IDE automation to remove System.out.println(..)s :-)
|
👍 after #2223 (comment) and commit squash |
navis
force-pushed
the
besteffort-split-locations
branch
from
ce1fd2d to
f03f7fb
Compare
|
removed log message and squashed |
|
@himanshug I think I have committer privilege but I don't know the rule. Can I commit my PR if some other committer approved it? I heard that in some projects, commit is allowed only by reviewer not by the author. |
|
@navis Please see http://druid.io/community/. You will need 2 +1s |
|
@fjy With two +1, then any committer including the author can commit it? |
|
@navis general rule is two +1's other than author. |
5 participants
Currently
DatasourceInputSplitdoes not provide server locations. This is best effort try to find for them.