Regarding the first question, it has to be used on scanners because
the paging is done at the row level, not cell level.

One example usage I can think of, the unit tests:
https://github.com/apache/hbase/blob/trunk/src/test/java/org/apache/hadoop/hbase/filter/TestFilter.java#L193

Regarding your second question... I'm not sure what exactly the
question is. Could you rephrase? Thanks.

J-D

On Mon, Dec 6, 2010 at 2:17 AM, 梁景明 <[email protected]> wrote:
> hi, is there any examples for pagefilter?
> here is my app case.
> person A and his friends
>
> id         column      value
> A         firend:B      B
> A         firend:C      C
> A         firend:D      D
> ................
> A         firend:Z      Z
>
> and i want to get A's friends from D to G
>
> as index from 2 to 5
>
> 5-2=3 so i think maybe set new PageFilter(3) ,but all results are output
> ---------------------------------------------------------------
>        Get get = new Get("A".getBytes());
>        get.addColumn("firend".getBytes());
>        get.setFilter(new PageFilter(3));
>        table.get(get);
>        Result r = table.get(get);
>       List <KeyValue> list=r.list();
>        for(KeyValue kv:list){
>            System.out.println(Bytes.toString(kv.getValue()));
>        }
>  ------------------------------------------------------------
> how to use it ?
>
> and more question ,some function what i need is some array use case.
>
> as
> previous,next,get(index), data(from x to y) ,size
>
> is there any new design for array in hbase?
>
> as desgin
> A         firend:B      B
>
> i get size from Result.list().size(), but if there is a bigtable ,i double
> it done well in memory?
>
> previous
> i scan the record current , mark the previous as result , when get current
> break,and get  previous one . but it's a bad idea which record is last one.
>
> next
> i scan the record current , when get current, mark the next as result , and
> get  next one . but it's also a bad idea which record is last one.
>
> get(index), data(from x to y)
> now  i have no idea? any suggestion?
>
> thanks verymuch.
>

Reply via email to