You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hbase.apache.org by johnson hong <ho...@goodhope.net> on 2010/01/19 03:32:19 UTC
column can be got but can not be mapped
There are abount 2,000,000 rows in HTable.
I have deleted some rows(column value is wrong,so delete it) with timestamp
in currentTime nanos,
then,in another mapReduce,I put new rows(rowKey is same as above ) with
timestamp in currentTime millis,
all the new rows do live in HTable because I can get them by hbase shell,but
the problem is :when I map this HTable again,they are never be scanned.
--
View this message in context: http://old.nabble.com/column-can-be-got-but-can-not-be-mapped-tp27220186p27220186.html
Sent from the HBase User mailing list archive at Nabble.com.
Re: column can be got but can not be mapped
Posted by johnson hong <ho...@goodhope.net>.
Thank you for your reply,I will run major compaction and try again soon.
Jean-Daniel Cryans-2 wrote:
>
> The problem here is you deleted something in the future so everything
> before that is also considered as being deleted. If you run a major
> compaction over that table it will all be cleaned (or if you don't do
> anything it will happen in less than 24 hours).
>
> Bottom line, it's better to always stay in the same time frame or
> weird stuff happens.
>
> J-D
>
> On Mon, Jan 18, 2010 at 6:32 PM, johnson hong
> <ho...@goodhope.net> wrote:
>>
>> There are abount 2,000,000 rows in HTable.
>> I have deleted some rows(column value is wrong,so delete it) with
>> timestamp
>> in currentTime nanos,
>> then,in another mapReduce,I put new rows(rowKey is same as above ) with
>> timestamp in currentTime millis,
>> all the new rows do live in HTable because I can get them by hbase
>> shell,but
>> the problem is :when I map this HTable again,they are never be scanned.
>> --
>> View this message in context:
>> http://old.nabble.com/column-can-be-got-but-can-not-be-mapped-tp27220186p27220186.html
>> Sent from the HBase User mailing list archive at Nabble.com.
>>
>>
>
>
--
View this message in context: http://old.nabble.com/column-can-be-got-but-can-not-be-mapped-tp27220186p27220858.html
Sent from the HBase User mailing list archive at Nabble.com.
Re: column can be got but can not be mapped
Posted by Jean-Daniel Cryans <jd...@apache.org>.
The problem here is you deleted something in the future so everything
before that is also considered as being deleted. If you run a major
compaction over that table it will all be cleaned (or if you don't do
anything it will happen in less than 24 hours).
Bottom line, it's better to always stay in the same time frame or
weird stuff happens.
J-D
On Mon, Jan 18, 2010 at 6:32 PM, johnson hong
<ho...@goodhope.net> wrote:
>
> There are abount 2,000,000 rows in HTable.
> I have deleted some rows(column value is wrong,so delete it) with timestamp
> in currentTime nanos,
> then,in another mapReduce,I put new rows(rowKey is same as above ) with
> timestamp in currentTime millis,
> all the new rows do live in HTable because I can get them by hbase shell,but
> the problem is :when I map this HTable again,they are never be scanned.
> --
> View this message in context: http://old.nabble.com/column-can-be-got-but-can-not-be-mapped-tp27220186p27220186.html
> Sent from the HBase User mailing list archive at Nabble.com.
>
>