Solr DataImportHandler foreign tables

Question

i've build a solr-index like here: https://wiki.apache.org/solr/DataImportHandler

I used something like this in my data-config.xml:

<entity name="article" pk="id" query="Select * from msp_article">
            <field column="id" name="id"/>
            <field column="title" name="docTitle"/>
            <field column="orgRef" name="ref" />
            <entity name="text" query="select content from msp_articleText where fk_articleId='${article.id}'">
                <field column="content" name="textContent"/>
            </entity>
        </entity>

When look in my db and test the keys, i get in the msp_articleText 3 rows, but in the solr-index its only the first row.

What's wrong here? Please help

Answer 1

Try to have a single query using the join like below

Select * from msp_article ma, msp_articleText mat where mat.fk_articleId = ma.msp_article.id

and then have the fields like

<field column="id" name="id"/>
<field column="title" name="docTitle"/>
<field column="orgRef" name="ref" />
<field column="content" name="textContent"/>

It should be like :

<entity name="article" pk="id" query="Select * from msp_article , msp_articleText where fk_articleId = msp_article.id">
            <field column="id" name="id"/>
            <field column="title" name="docTitle"/>
            <field column="orgRef" name="ref" />
            <field column="content" name="textContent"/>
</entity>

You can build your own query as you want, I just wanted to convey that instead of 2 entities go for one where you will have a join query.

Answer 2

Thanks, but can this work? When i try it i get 3 rows with same id. The id is the pk for solr.

For the moment i use SELECT GROUP_CONCAT(content) because the content is to big, i put it in a procedure

CREATE PROCEDURE `msp_bla`(IN con CHAR(20))
BEGIN
    SET SESSION group_concat_max_len = 1000000;
    SELECT GROUP_CONCAT(content) as content from msp_articleText WHERE fk_articleId = con group by fk_articleId;
END

And in my data-config.xml i use

<entity name="text" query="call msp_bla('${article.id}')">

It works, but im not happy with that.