 |
Nathan Marz |
at Jan 8, 2012 at 6:00 am
|
⇧ | 
| |
This is currently not possible, as Cascading does not provide that
flexibility. The only way to do this right now is to break up the query
into multiple stages with the intermediate outputs sent to temporary paths.
cascalog-checkpoint helps a lot with doing stuff like that (see
http://sritchie.github.com/2011/11/15/introducing-cascalogcontrib.html for
more details).
I opened up an issue for this since it would be useful:
https://github.com/nathanmarz/cascalog/issues/43On Sat, Jan 7, 2012 at 1:05 AM, Artem Boytsov wrote:
Hello, Nathan et al,
Is it possible to specify hadoop configuration parameters (e.g.
mapred.map.tasks etc.) for each _query_? with-job-conf does the job, but if
I have nested queries, they sometimes need different configuration.
Thanks,
Artem.
--
Twitter: @nathanmarz
http://nathanmarz.com
