Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Accumulo, mail # user - How Do I Add Split From Shell When Split Value Has Zero (\x00) Byte?


Copy link to this message
-
Re: How Do I Add Split From Shell When Split Value Has Zero (\x00) Byte?
Keith Turner 2013-04-11, 21:47
| think was fixed in 1.5 in changes related to ACCUMULO-596, r1339751

On Thu, Apr 11, 2013 at 3:15 PM, David Medinets
<[EMAIL PROTECTED]> wrote:
> I tried two different ways to represent the zero bye. One was \x00 and one
> was \0x00. The second representation \0x00 caused the BadArgumentException.
> The first representation was accepted by the shell but either was not
> displayed correctly by the getsplits command or the zero byte was ignored by
> the addsplits command.
>
>
> On Thu, Apr 11, 2013 at 2:03 PM, Keith Turner <[EMAIL PROTECTED]> wrote:
>>
>> On Thu, Apr 11, 2013 at 12:46 PM, David Medinets
>> <[EMAIL PROTECTED]> wrote:
>> > In 1.4.1, using \0x00 results in an error message: BadArgumentException:
>> > can
>> > only escape single quotes, double quotes, the space character, the
>> > backslash, and hex input.
>>
>> Ok, so what was in the email was a mistyped version of what you actually
>> ran?
>>
>> >
>> >
>> > On Thu, Apr 11, 2013 at 10:53 AM, Keith Turner <[EMAIL PROTECTED]> wrote:
>> >>
>> >> On Thu, Apr 11, 2013 at 10:42 AM, Keith Turner <[EMAIL PROTECTED]>
>> >> wrote:
>> >> > On Thu, Apr 11, 2013 at 10:38 AM, David Medinets
>> >> > <[EMAIL PROTECTED]> wrote:
>> >> >> I tried that addsplits with a zero byte and ran into something
>> >> >> unexpected:
>> >> >>
>> >> >> createtable zerotest
>> >> >> insert AA\x00BB cf cq value
>> >> >> insert AA\x00CC cf cq value
>> >> >> insert AA\x01CC cf cq value
>> >> >> addsplits AA\0x00CC
>> >>
>> >> just noticed this should \0x00 be \x00?
>> >>
>> >> >> getsplits
>> >> >> AACC <--- the zero byte is not displayed.
>> >> >>
>> >> >> This seems like a bug. If anyone can verify, I'll create a jira
>> >> >> ticket.
>> >> >
>> >> > that does seem like a bug.  Either the split was not added correctly,
>> >> > or its not being displayed correctly.  Assuming it was added
>> >> > correctly, the output of getsplits may not use the same code that
>> >> > scans do to display unprintable bytes.
>> >> >
>> >> >>
>> >> >>
>> >> >> On Thu, Apr 11, 2013 at 9:44 AM, Josh Elser <[EMAIL PROTECTED]>
>> >> >> wrote:
>> >> >>>
>> >> >>> AFAIK you should be able to run "addsplits AAA\x00CCC" in the
>> >> >>> shell.
>> >> >>> Did
>> >> >>> that not work?
>> >> >>>
>> >> >>>
>> >> >>> On 4/11/13 9:24 AM, David Medinets wrote:
>> >> >>>>
>> >> >>>> I have rows that look like this when displayed through the shell:
>> >> >>>>
>> >> >>>> AAA\x00BBB
>> >> >>>> AAA\x00BBB
>> >> >>>> AAA\x00CCC
>> >> >>>> AAA\x00CCC
>> >> >>>>
>> >> >>>> I would like to manually add a split for AAA\x00CCC. I know the
>> >> >>>> shell
>> >> >>>> has
>> >> >>>> some restrictions on how some characters are handled so I thought
>> >> >>>> I'd
>> >> >>>> ask
>> >> >>>> before trying.
>> >> >>>
>> >> >>>
>> >> >>
>> >
>> >
>
>