pig UDF is lacking XMLLoader. Plan to add the XMLLoader

Hi All,

We are planning to add the XMLLoader UDF in the piggybank repository.

Here is the proposal with the user docs :-

The load function to load the XML file
This will implements the LoadFunc interface which is used to parse records
from a dataset.
This takes a xmlTag as the arg which it will use to split the inputdataset into
multiple records.

For example if the input xml (input.xml) is like this
<configuration>
<property>
<name> foobar </name>
<value> barfoo </value>
</property>
<ignoreProperty>
<name> foo </name>
</ignoreProperty>
<property>
<name> justname </name>
</property>
</configuration>

And your pig script is like this

--load the jar files
register loader.jar;
– load the dataset using XMLLoader
– A is the bag containing the tuple which contains one atom i.e doc see output
A = load '/user/aloks/pig/input.xml using loader.XMLLoader('property') as (doc:chararray);
--dump the result
dump A;

Then you will get the output

(<property>
<name> foobar </name>
<value> barfoo </value>
</property>)
(<property>
<name> justname </name>
</property>)

Where each () indicate one record