hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Henry JunYoung KIM <henry.jy...@gmail.com>
Subject Re: ways to make orders when it puts
Date Fri, 05 Oct 2012 01:58:28 GMT
yes, this needs for our indexer for datas.

I mean that hbase need to store some kinds of data list based on entry time and then by indexer,
It will try to search new data list by a start-key and a limit count.

for easy understanding, 

If I used a timestamp row key, 

ts    data
1       D1
2       D2
3       D3
4       D4

an indexer has done its job until ts 3. 
but at this point, an one of threads stored its data into a row key ts 2.

this situation will give rise to lose of ts 2 for an indexer even if it should be indexed.

because of this, I gave a questions to store datas based on entry time. 

thanks for your concerns.

2012. 10. 5., 오전 1:06, Michael Segel <michael_segel@hotmail.com> 작성:

> I took it that the OP wants to store the rows A1->A3 in the order in which they came
in. So  It could be A3,A1,A2  as an example. 
> So to do this you end up prefixing the rowkey with a timestamp or something. 
> This is not a good idea, and I was curious as to why the order of entry was important
to the OP.
> On Oct 4, 2012, at 10:54 AM, Dan Han <dannahan2008@gmail.com> wrote:
>> I am not sure I hit your questoin, but if the data is not stored as what
>> you expect,
>> I guess it might be the problem of row key.
>> As we all know, the row key is sorted in a lexicographic order in HBase.
>> For example, 10 is before 9. So if your row key includes 1 ... 10,
>> it is neccessory to format the single letter by adding "0".
>> Best Wishes
>> Dan Han
>> On Thu, Oct 4, 2012 at 9:19 AM, Michael Segel <michael_segel@hotmail.com>wrote:
>>> Silly question. Why do you care how your data is being stored?
>>> Does it matter if the data is stored in rows where A1,A2, A3 are the order
>>> of the keys, or
>>> if its A3,A1,A2 ?
>>> If you say that you want to store the rows in order based on entry time,
>>> you're going to also have to deal with a little nasty problem of hot
>>> spotting along with your regions being only half full post spilt.
>>> On Oct 4, 2012, at 3:54 AM, JUN YOUNG KIM <juneng603@me.com> wrote:
>>>> hi, hbase users.
>>>> I am wondering how we can make orders when we put under multiple threads.
>>>> I mean that
>>>> threads are working like this
>>>> thread1 puts A1 (rowkey)
>>>> thread2 puts A2
>>>> thread3 puts A3
>>>> by unexpected working time order,
>>>> thread1 puts earlier than thread2.
>>>> thread3 puts earlier than thread1.
>>>> yes, I know that hbase will store it in-order like A1 -> A2 -> A3
>>>> but, how could I store my datas by write-times like A3 -> A1 -> A2
>>>> If I could insert timestamp value before A#, the situations I described
>>> could be also happened.
>>>> any ideas??
>>>> (you can change row key structure if you can satisfy conditions I want
>>> to archive.)
>>>> thanks for your concerns.

View raw message