DataImportHandler: Javascript transformer for splitting field-values

[Steffen <el...@gmail.com>]

Hi everyone,
I'm currently working with the nightly build of Solr (solr-2008-11-17)
and trying to figure out how to transform a row-object with Javascript
to include multiple values (in a single multivalued field). When I try
something like this as a transformer:
function splitTerms(row) {
                        //each term should be duplicated into count field-values
                        //dummy-code to show the idea
                	row.put('terms',['term','term','term']);
                	return row;
}
[...]
<entity name="searchfeedback" pk="id" transformer="script:splitTerms"
query="SELECT term,count FROM termtable WHERE id=${parent.id}" />

The DataImportHandler debugger returns:
<arr>
  <str>
    sun.org.mozilla.javascript.internal.NativeArray:sun.org.mozilla.javascript.internal.NativeArray@6a4268
  </str>
</arr>
What it *should* return:
<arr>
  <str>term</str>
  <str>term</str>
  <str>term</str>
</arr>

So, what am I doing wrong? My transformer will be invoked multiple
times from a MySQL-Query and in turn has to insert multiple values to
the same field during each invocation. It should do something similar
to the RegexTransformer (field splitBy)... is that possible? Right now
I have to use a workaround that includes the term-duplication on the
database sides, which is kinda ugly if a term has to be duplicated a
lot.
Greetings,
Steffen

Re: DataImportHandler: Javascript transformer for splitting field-values

[Noble Paul നോബിള്‍ नोब्ळ् <no...@gmail.com>]

unfortunately native JS objects are not handled by the ScriptTransformer yet.

but what you can do in the script is create a new
java.util.ArrayList() and add each item into that .

some thing like
var jsarr = ['term','term','term']
var arr = new java.util.ArrayList();
for each in jsarr... arr.add(item)
row.put('terms',arr);

On Wed, Nov 19, 2008 at 9:03 PM, Steffen <el...@gmail.com> wrote:
> Hi everyone,
> I'm currently working with the nightly build of Solr (solr-2008-11-17)
> and trying to figure out how to transform a row-object with Javascript
> to include multiple values (in a single multivalued field). When I try
> something like this as a transformer:
> function splitTerms(row) {
>                        //each term should be duplicated into count field-values
>                        //dummy-code to show the idea
>                        row.put('terms',['term','term','term']);
>                        return row;
> }
> [...]
> <entity name="searchfeedback" pk="id" transformer="script:splitTerms"
> query="SELECT term,count FROM termtable WHERE id=${parent.id}" />
>
> The DataImportHandler debugger returns:
> <arr>
>  <str>
>    sun.org.mozilla.javascript.internal.NativeArray:sun.org.mozilla.javascript.internal.NativeArray@6a4268
>  </str>
> </arr>
> What it *should* return:
> <arr>
>  <str>term</str>
>  <str>term</str>
>  <str>term</str>
> </arr>
>
> So, what am I doing wrong? My transformer will be invoked multiple
> times from a MySQL-Query and in turn has to insert multiple values to
> the same field during each invocation. It should do something similar
> to the RegexTransformer (field splitBy)... is that possible? Right now
> I have to use a workaround that includes the term-duplication on the
> database sides, which is kinda ugly if a term has to be duplicated a
> lot.
> Greetings,
> Steffen
>



-- 
--Noble Paul