I have some questions about the way Hbase returns results. When we use
HTable getScanner(Scan), it looks like it only retrieves 1 row per next()
call which is different from how jdbc returns the resultset. If I set the
Filter on the Scan, could it return a set of rows in one connection call? It
looks like the ClientScanner does not make use of the Filter. Only the Scan
uses the Filter in the readFields method. Please correct me if i am wrong
about this.

Thank you so much.

The API shows one row per next() call but the number of rows fetched per RPC can be configured much higher with Scan.setCaching().

Filters are basically just server-side predicates that will dictate which rows/columns/values will be returned to the client.  This does not relate to the number of rows sent per RPC.  See http://hbase.apache.org/docs/current/api/org/apache/hadoop/hbase/filter/package-summary.html for more information about filters.

JG

> -----Original Message-----
> From: Weishung Chung [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, January 18, 2011 8:31 AM
> To: [EMAIL PROTECTED]
> Subject: Scan with Filter
>
> I have some questions about the way Hbase returns results. When we use
> HTable getScanner(Scan), it looks like it only retrieves 1 row per next() call
> which is different from how jdbc returns the resultset. If I set the Filter on
> the Scan, could it return a set of rows in one connection call? It looks like the
> ClientScanner does not make use of the Filter. Only the Scan uses the Filter in
> the readFields method. Please correct me if i am wrong about this.
>
> Thank you so much.

Thank you :)

On Tue, Jan 18, 2011 at 12:49 PM, Jonathan Gray <[EMAIL PROTECTED]> wrote:

> The API shows one row per next() call but the number of rows fetched per
> RPC can be configured much higher with Scan.setCaching().
>
> Filters are basically just server-side predicates that will dictate which
> rows/columns/values will be returned to the client.  This does not relate to
> the number of rows sent per RPC.  See
> http://hbase.apache.org/docs/current/api/org/apache/hadoop/hbase/filter/package-summary.htmlfor more information about filters.
>
> JG
>
> > -----Original Message-----
> > From: Weishung Chung [mailto:[EMAIL PROTECTED]]
> > Sent: Tuesday, January 18, 2011 8:31 AM
> > To: [EMAIL PROTECTED]
> > Subject: Scan with Filter
> >
> > I have some questions about the way Hbase returns results. When we use
> > HTable getScanner(Scan), it looks like it only retrieves 1 row per next()
> call
> > which is different from how jdbc returns the resultset. If I set the
> Filter on
> > the Scan, could it return a set of rows in one connection call? It looks
> like the
> > ClientScanner does not make use of the Filter. Only the Scan uses the
> Filter in
> > the readFields method. Please correct me if i am wrong about this.
> >
> > Thank you so much.
>

If I expect the valid scan result to be more than 10k (just a big number),
how are we going to set this number properly? -- if I set it to 1k, then
there will be 10 RPC round trips.
On Tue, Jan 18, 2011 at 1:24 PM, Weishung Chung <[EMAIL PROTECTED]> wrote:

> Thank you :)
>
> On Tue, Jan 18, 2011 at 12:49 PM, Jonathan Gray <[EMAIL PROTECTED]> wrote:
>
> > The API shows one row per next() call but the number of rows fetched per
> > RPC can be configured much higher with Scan.setCaching().
> >
> > Filters are basically just server-side predicates that will dictate which
> > rows/columns/values will be returned to the client.  This does not relate
> to
> > the number of rows sent per RPC.  See
> >
> http://hbase.apache.org/docs/current/api/org/apache/hadoop/hbase/filter/package-summary.htmlformore information about filters.
> >
> > JG
> >
> > > -----Original Message-----
> > > From: Weishung Chung [mailto:[EMAIL PROTECTED]]
> > > Sent: Tuesday, January 18, 2011 8:31 AM
> > > To: [EMAIL PROTECTED]
> > > Subject: Scan with Filter
> > >
> > > I have some questions about the way Hbase returns results. When we use
> > > HTable getScanner(Scan), it looks like it only retrieves 1 row per
> next()
> > call
> > > which is different from how jdbc returns the resultset. If I set the
> > Filter on
> > > the Scan, could it return a set of rows in one connection call? It
> looks
> > like the
> > > ClientScanner does not make use of the Filter. Only the Scan uses the
> > Filter in
> > > the readFields method. Please correct me if i am wrong about this.
> > >
> > > Thank you so much.
> >
>

--
--Sean

Hi,
I have a question about Scan.

 *Scan s = new Scan();
 s.setFilter(f);
 ResultScanner scanner = _table.getScanner(s); *

I want to know the operation about filter was done at server-side  or
client-side?

the data was sent back to client-side before filtering or the data was
filtered then sent to client ?
Thanks
--
Sindy