You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by petri koski <mo...@gmail.com> on 2012/10/31 16:19:32 UTC
Hbase 0,9X + Hadoop Columns return
I am totally stuck here.
I have table which is called url and where are p, and i and s families.
Table url have 8300 rows.
Those rows are inserted like
key:xxxxx columnfamily:p: value:<webpage content>
Now, when I do scan in Hadoop, I add right columnFamily p, and try to
process all those 8300 rows in one map phase (I use multithreaderMapper
-patch, and use synchronize inside mapper etc.) I get only 566 rows (Map
input) and NOT those 8300 rows I am expecting to process.
What could possible be wrong ? I process those inputs in my Mapper as:
public void map(ImmutableBytesWritable row, Result values,Context context)
throws IOException {
for(KeyValue kv : values.raw()){
String i = new String(kv.getRow());
String p = new String(kv.getValue());
do something with p ..
savecontent(save processed things to table url family ilinks)
}
}
To get and idea .. What could be the reason ? I dont use any start or stop
rows ..
Yours,
Petri
Re: Hbase 0,9X + Hadoop Columns return
Posted by Jean-Daniel Cryans <jd...@apache.org>.
issues@ is the wrong mailing list for this, I'm putting it in BCC and
replying now to user@
First guess that comes in mind is that not all your rows have data in
p:, if it's really the only family you are scanning in your MR.
Trying doing a scan in the shell in p: and see what comes out.
J-D
On Wed, Oct 31, 2012 at 8:19 AM, petri koski <mo...@gmail.com> wrote:
> I am totally stuck here.
>
> I have table which is called url and where are p, and i and s families.
>
> Table url have 8300 rows.
>
> Those rows are inserted like
>
> key:xxxxx columnfamily:p: value:<webpage content>
>
> Now, when I do scan in Hadoop, I add right columnFamily p, and try to
> process all those 8300 rows in one map phase (I use multithreaderMapper
> -patch, and use synchronize inside mapper etc.) I get only 566 rows (Map
> input) and NOT those 8300 rows I am expecting to process.
>
> What could possible be wrong ? I process those inputs in my Mapper as:
>
> public void map(ImmutableBytesWritable row, Result values,Context context)
> throws IOException {
>
> for(KeyValue kv : values.raw()){
>
> String i = new String(kv.getRow());
> String p = new String(kv.getValue());
>
> do something with p ..
>
> savecontent(save processed things to table url family ilinks)
>
> }
> }
>
> To get and idea .. What could be the reason ? I dont use any start or stop
> rows ..
>
> Yours,
>
> Petri
Re: Hbase 0,9X + Hadoop Columns return
Posted by Jean-Daniel Cryans <jd...@apache.org>.
issues@ is the wrong mailing list for this, I'm putting it in BCC and
replying now to user@
First guess that comes in mind is that not all your rows have data in
p:, if it's really the only family you are scanning in your MR.
Trying doing a scan in the shell in p: and see what comes out.
J-D
On Wed, Oct 31, 2012 at 8:19 AM, petri koski <mo...@gmail.com> wrote:
> I am totally stuck here.
>
> I have table which is called url and where are p, and i and s families.
>
> Table url have 8300 rows.
>
> Those rows are inserted like
>
> key:xxxxx columnfamily:p: value:<webpage content>
>
> Now, when I do scan in Hadoop, I add right columnFamily p, and try to
> process all those 8300 rows in one map phase (I use multithreaderMapper
> -patch, and use synchronize inside mapper etc.) I get only 566 rows (Map
> input) and NOT those 8300 rows I am expecting to process.
>
> What could possible be wrong ? I process those inputs in my Mapper as:
>
> public void map(ImmutableBytesWritable row, Result values,Context context)
> throws IOException {
>
> for(KeyValue kv : values.raw()){
>
> String i = new String(kv.getRow());
> String p = new String(kv.getValue());
>
> do something with p ..
>
> savecontent(save processed things to table url family ilinks)
>
> }
> }
>
> To get and idea .. What could be the reason ? I dont use any start or stop
> rows ..
>
> Yours,
>
> Petri