My schema.xml is splitting product name and then uses RemoveDuplicate to remove duplicated words after split.
<fieldType name="type_name" class="solr.TextField">
<analyzer type="index">
<tokenizer class="solr.PatternTokenizerFactory" pattern="\|| " />
<filter class="solr.LowerCaseFilterFactory"/>
<filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
And in query analyzer I see that RemoveDuplicatesTokenFilterFactory did absolutely nothing to duplicated words. Why?

If you read Wiki you will see that it only removes duplicates at the same position, which is not the case here.