Yeikel

I tested the query building and it works as expected. You did an amazing job , thank you! The only question that remains is the analyzer.

I believe something it is not working as expected. Given my config : ``` lucenerdd { // Name of analyzer as it is under Lucene's package org.apache.lucene.analysis.XX analyzer.name = "StandardAnalyzer"...

My assumption is incorrect. The character "." is not part of the characters removed by the Standard Analyzer.

I believe this issue can be considered done unless we find any other bug. #156 addresses the possibility of custom analyzers so it is probably better to move the conversation...

> It is currently available under `0.3.6-SNAPSHOT`. Feel free to test. > > One of my concerns is that since we specify the `Query` now, there might be a mismatched...

Many other types do not support query as well. For example : ``` def dedup[T1: ClassTag](searchQueryGen: T1 => String, topK: Int = DefaultTopK, linkerMethod: String = getLinkerMethod) : RDD[(T1, Array[SparkScoreDoc])]...

> Yes only the block linkage methods support Query type for now. > […](#) > On Tue, 19 Mar 2019, 15:59 Yeikel, ***@***.***> wrote: Many other types do not support...

I tried to enhance it but I am not sure why it works sometimes and why it is producing a serialization error for me . How were you able to...

So if `linker.method = "collectbroadcast"` , I see this exception. When `linker.method = "cartesian"` it works fine. I still don't understand how it is not failing for the tests you...

I added a little bit more of support in this commit https://github.com/yeikel/spark-lucenerdd/commit/3e408c7fbd884aaeaae14a9add0c4151dd904017. Could you please review it?