Solr 4-索引发布的文本文件

Question

I'm trying to create a field called "sku" - which is indexed with the following analyzer:

<fieldType name="sku" class="solr.TextField">
   <analyzer>
  <tokenizer class="solr.PatternTokenizerFactory" pattern="(SKU|Part(\sNumber)?):?\s(\[0-9-\]+)" group="3"/>
</analyzer>
 </fieldType>

This is from reading the documentation here http://lucidworks.lucidimagination.com/display/solr/Tokenizers#Tokenizers-RegularExpressionPatternTokenizer

I already have a Java program that is posting to the solr server succesfully, however it is not grabbing the sku out of any files, and indexing them. Here is my Java code:

ContentStreamUpdateRequest up = new ContentStreamUpdateRequest(
                "/update/extract");
        up.addFile(arg0, arg0.getName());

        up.setParam("literal.id", arg0.getName());
        up.setParam("uprefix", "attr_");
        up.setParam("fmap.content", "attr_content");

        up.setAction(AbstractUpdateRequest.ACTION.COMMIT, true, true);

        server.request(up);

Any help appreciated.

I understand I can parse the text files myself and extract the SKU and post them in the parameters to the server, but I thought Solr could do this for me?

Answer 1

It is hard to tell what is going on, because there is a several steps in the middle.

For example, what's your schema.xml definition. Is it definitely using sku as its type (and not say string). Then, what's the field name (attr_sku?) and does the extract handler mapping actually maps to it properly? The extract handler usually sends metadata as individual fields and then all file content as one big long field. Is sku somewhere in metadata?

I would do a copyField into something non-processing and see whether the content actually makes it into Solr field. Then, I would start troubleshooting the regex itself.