hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ted Yu <yuzhih...@gmail.com>
Subject Re: Delete rowKey in hexadecimal array bytes.
Date Tue, 10 Jun 2014 12:35:15 GMT
Can you pastebin your junit test ?

Thanks

On Jun 10, 2014, at 3:07 AM, gortiz <gortiz@pragsis.com> wrote:

> I think we are in different points :).
> 
> The problem is when the comparation of the keys when I use Hex.decode
> 
> I emit keyValues on this way and works, but it spends double of memory to store keys.
> 
> rowKey = Bytes.toBytes(DigestUtils.*shaHex*(pOutput.getRow()));
> KeyValue kv =
>                    new KeyValue(Bytes.toBytes(pOutput.getRow()), family, null, ts, type,
null);
> 
> If I use for rowKey:
>    rowKey = DigestUtils.*sha*(pOutput.getRow());
> it doesn't work, I don't know why, since it's an byte array. But, I coded some Junit
test and it never deletes the keys.
> 
> 
> 
> 
> On 09/06/14 13:43, Ted Yu wrote:
>> Decode rowkey has timestamp.
>> KeyValue has timestamp field.
>> 
>> Do these two timestamps carry the same value ?
>> 
>> Cheers
>> 
>> On Jun 9, 2014, at 2:02 AM, Guillermo Ortiz <konstt2000@gmail.com> wrote:
>> 
>>> Hi,
>>> 
>>> I'm generating key with SHA1, as it's a hex representation after generating
>>> the keys, I use Hex.decode to save memory since I could store them in half
>>> space.
>>> 
>>> I have a MapReduce process which deletes some of these keys, the problem
>>> it's that it's that when I try to delete them, but I don't get it. If I
>>> don't do the parse to Hex, it works.
>>> 
>>> So, For example, I put the keys in SHA like
>>> b343664e210e7a7abff3625a005e65e2b0d4616 works, but if I parse this key with
>>> Hex.decode to *\xB3CfN!\x0Ezz\xBF\xF3bZ\x00^e\xE2\xB0\ *column=l:dd,
>>> timestamp=1384317115000 .... it doesn't.
>>> 
>>> I have been checked the code a lot but I think it's right, plus, if I
>>> comments the decode to Hex it works.
>>> 
>>> Any clue about it? is there any problem with I am trying to??
> 
> 
> -- 
> *Guillermo Ortiz*
> /Big Data Developer/
> 
> Telf.: +34 917 680 490
> Fax: +34 913 833 301
> C/ Manuel Tovar, 49-53 - 28034 Madrid - Spain
> 
> _http://www.bidoop.es_
> 

Mime
View raw message