某因幡 2012-08-27, 10:42
Ted Yu 2012-08-27, 13:23
某因幡 2012-08-28, 03:50
某因幡 2012-08-29, 01:54
-Re: PIG with HBase Coprocessors
yuzhihong@... 2012-08-29, 02:16
Did you find some clue from region server logs ?
You can pastebin snippet and show the link here.
On Aug 28, 2012, at 6:54 PM, 某因幡 <[EMAIL PROTECTED]> wrote:
> 2012/8/28 某因幡 <[EMAIL PROTECTED]>:
>> Thanks for your quick reply.
>> The co-processor looks like:
>> public void postGet(final ObserverContext<RegionCoprocessorEnvironment> e,
>> final Get get, final List<KeyValue> results)
>> if table is X
>> get some columns from table Y
>> add these columns to results
>> And similar for postScannerNext().
>> This works in HBase JRuby shell using get/scan plus heavy performance loss.
>> And the versions are:
>> Hadoop 1.0.3
>> HBase 0.92.1
>> Pig 0.9.2
>> 2012/8/27 Ted Yu <[EMAIL PROTECTED]>:
>>> Allow me to refer to previous discussion:
>>> http://mail-archives.apache.org/mod_mbox/hbase-user/201203.mbox/%3CCABsY1jQ8+OiLh7SYkXZ8iO=NosY8khz7iYs+[EMAIL PROTECTED]%3E
>>> If the above doesn't answer your question, please give us more details
>>> about the versions of HBase and PIG you're using.
>>> On Mon, Aug 27, 2012 at 3:42 AM, 某因幡 <[EMAIL PROTECTED]> wrote:
>>>> I've created a co-processor which will insert more columns in the
>>>> result by overriding preGet and postScannerNext.
>>>> And it is registered as a system co-processor, works with hbase shell,
>>>> both get and scan.
>>>> When I tried to access those columns in PIG, it simply returns null
>>>> for the column.
>>>> Is there something more to be done?
>>>> language: Chinese, Japanese, English
>> language: Chinese, Japanese, English
> language: Chinese, Japanese, English