|
|
-
Counters from a PIG Job
rakesh kothari 2010-10-07, 23:48
Hi,
Couple of Questions:
1. What's the best way to get "Counters" out of a Pig Job execution ? (e.g. Counters object exposed by "org.apache.mapreduce.hadoop.Job")
2. Is there anything special needs to be done to have PigStats taken (i.e. any property etc ) ? I get empty PigStats when I run my job in local mode.
Thanks, -Rakesh
+
rakesh kothari 2010-10-07, 23:48
-
Re: Counters from a PIG Job
Dmitriy Ryaboy 2010-10-08, 03:10
What version of pig are you using?
On Thu, Oct 7, 2010 at 4:48 PM, rakesh kothari <[EMAIL PROTECTED]>wrote:
> > Hi, > > Couple of Questions: > > 1. What's the best way to get "Counters" out of a Pig Job execution ? (e.g. > Counters object exposed by "org.apache.mapreduce.hadoop.Job") > > 2. Is there anything special needs to be done to have PigStats taken (i.e. > any property etc ) ? I get empty PigStats when I run my job in local mode. > > Thanks, > -Rakesh > >
+
Dmitriy Ryaboy 2010-10-08, 03:10
-
Re: Counters from a PIG Job
Ashutosh Chauhan 2010-10-08, 03:34
Assuming you are either on trunk / 0.8 branch. Way to get counters is through PigStats. But, in local mode Hadoop doesnt provide any counters to Pig, so they arent available. You will get all the counters in map-reduce mode via PigStats.
Hope it helps, Ashutosh
On Thu, Oct 7, 2010 at 20:10, Dmitriy Ryaboy <[EMAIL PROTECTED]> wrote: > What version of pig are you using? > > On Thu, Oct 7, 2010 at 4:48 PM, rakesh kothari <[EMAIL PROTECTED]>wrote: > >> >> Hi, >> >> Couple of Questions: >> >> 1. What's the best way to get "Counters" out of a Pig Job execution ? (e.g. >> Counters object exposed by "org.apache.mapreduce.hadoop.Job") >> >> 2. Is there anything special needs to be done to have PigStats taken (i.e. >> any property etc ) ? I get empty PigStats when I run my job in local mode. >> >> Thanks, >> -Rakesh >> >> >
+
Ashutosh Chauhan 2010-10-08, 03:34
-
RE: Counters from a PIG Job
rakesh kothari 2010-10-08, 20:25
I am using PIG 0.7 and I am able to get counters from PigStats when running the job in mapreduce mode.
Thanks, -Rakesh
> From: [EMAIL PROTECTED] > Date: Thu, 7 Oct 2010 20:34:08 -0700 > Subject: Re: Counters from a PIG Job > To: [EMAIL PROTECTED] > CC: [EMAIL PROTECTED] > > Assuming you are either on trunk / 0.8 branch. Way to get counters is > through PigStats. But, in local mode Hadoop doesnt provide any > counters to Pig, so they arent available. You will get all the > counters in map-reduce mode via PigStats. > > Hope it helps, > Ashutosh > > On Thu, Oct 7, 2010 at 20:10, Dmitriy Ryaboy <[EMAIL PROTECTED]> wrote: > > What version of pig are you using? > > > > On Thu, Oct 7, 2010 at 4:48 PM, rakesh kothari <[EMAIL PROTECTED]>wrote: > > > >> > >> Hi, > >> > >> Couple of Questions: > >> > >> 1. What's the best way to get "Counters" out of a Pig Job execution ? (e.g. > >> Counters object exposed by "org.apache.mapreduce.hadoop.Job") > >> > >> 2. Is there anything special needs to be done to have PigStats taken (i.e. > >> any property etc ) ? I get empty PigStats when I run my job in local mode. > >> > >> Thanks, > >> -Rakesh > >> > >> > >
+
rakesh kothari 2010-10-08, 20:25
|
|