Iinput path pointing on an hbase table and not a text file?

[Xine Jar <xi...@googlemail.com>]

Hallo,
I am using hadoop 0.19.1 and I was previously writing my application setting
a text file as an input path for the job.
*
Therefore the main four involved parameters were the following:*

public static class Map extends MapReduceBase implements
Mapper<LongWritable, *Text*, Text, Text> {
.......
  public void map(LongWritable key,*Text value*, OutputCollector<Text, Text>
output, Reporter reporter) throws IOException {
....}
}

public static void main(String[] args) throws Exception {
....
    conf.setInputFormat(*TextInputFormat*.class);
    FileInputFormat.setInputPaths(conf, *new Path(args[0]*));

....
}

*Question
*I would like to write a mapreduce application that reads from an hbase
table, summing some values etc,...and writing the output result in a
textfile. I have already created this hbase table (using version 0.19.3) I
need now to pass it's name as an input path to the job how can I do it?

I have red here and there and found out that the InputFormat shall be set to
"*TableInputFormat.class*" but what about the setInputPaths ?and the Value
class in the mapper?

Thank you,
CJ

Re: Iinput path pointing on an hbase table and not a text file?

[Xine Jar <xi...@googlemail.com>]

Thank you for the hints,
I have checked these two Java files and managed to make my own program
working.

Regards C.J.


On Fri, Aug 7, 2009 at 5:44 PM, stack <st...@duboce.net> wrote:

> Yeah, use TableInputFormat.  You tell it table name and column(s) you want
> to see in your mapper.  For how to set tablename, etc., see
> TableMapReduceUtil in the mapred package and then see RowCounter in the
> same
> package for an example that uses TableInputFormat, setting table name and
> columns read from the command-line via TableMapReduceUtil.
>
> St.Ack
>
>
> On Fri, Aug 7, 2009 at 7:41 AM, Xine Jar <xi...@googlemail.com> wrote:
>
> > Hallo,
> > I am using hadoop 0.19.1 and I was previously writing my application
> > setting
> > a text file as an input path for the job.
> > *
> > Therefore the main four involved parameters were the following:*
> >
> > public static class Map extends MapReduceBase implements
> > Mapper<LongWritable, *Text*, Text, Text> {
> > .......
> >  public void map(LongWritable key,*Text value*, OutputCollector<Text,
> Text>
> > output, Reporter reporter) throws IOException {
> > ....}
> > }
> >
> > public static void main(String[] args) throws Exception {
> > ....
> >    conf.setInputFormat(*TextInputFormat*.class);
> >    FileInputFormat.setInputPaths(conf, *new Path(args[0]*));
> >
> > ....
> > }
> >
> > *Question
> > *I would like to write a mapreduce application that reads from an hbase
> > table, summing some values etc,...and writing the output result in a
> > textfile. I have already created this hbase table (using version 0.19.3)
> I
> > need now to pass it's name as an input path to the job how can I do it?
> >
> > I have red here and there and found out that the InputFormat shall be set
> > to
> > "*TableInputFormat.class*" but what about the setInputPaths ?and the
> Value
> > class in the mapper?
> >
> > Thank you,
> > CJ
> >
>

Re: Iinput path pointing on an hbase table and not a text file?

[stack <st...@duboce.net>]

Yeah, use TableInputFormat.  You tell it table name and column(s) you want
to see in your mapper.  For how to set tablename, etc., see
TableMapReduceUtil in the mapred package and then see RowCounter in the same
package for an example that uses TableInputFormat, setting table name and
columns read from the command-line via TableMapReduceUtil.

St.Ack


On Fri, Aug 7, 2009 at 7:41 AM, Xine Jar <xi...@googlemail.com> wrote:

> Hallo,
> I am using hadoop 0.19.1 and I was previously writing my application
> setting
> a text file as an input path for the job.
> *
> Therefore the main four involved parameters were the following:*
>
> public static class Map extends MapReduceBase implements
> Mapper<LongWritable, *Text*, Text, Text> {
> .......
>  public void map(LongWritable key,*Text value*, OutputCollector<Text, Text>
> output, Reporter reporter) throws IOException {
> ....}
> }
>
> public static void main(String[] args) throws Exception {
> ....
>    conf.setInputFormat(*TextInputFormat*.class);
>    FileInputFormat.setInputPaths(conf, *new Path(args[0]*));
>
> ....
> }
>
> *Question
> *I would like to write a mapreduce application that reads from an hbase
> table, summing some values etc,...and writing the output result in a
> textfile. I have already created this hbase table (using version 0.19.3) I
> need now to pass it's name as an input path to the job how can I do it?
>
> I have red here and there and found out that the InputFormat shall be set
> to
> "*TableInputFormat.class*" but what about the setInputPaths ?and the Value
> class in the mapper?
>
> Thank you,
> CJ
>