ElasticSearch catenate_words -- only keep concatenated value - elasticsearch

Question

Following examples here: https://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-word-delimiter-graph-tokenfilter.html
Specifically the catenate_words option.
I would like to use this to concatenate words that I can then use in a phrase query before and after the concatenated word, but the word parts prevent this.
For example, their example is this:
super-duper-xl → [ superduperxl, super, duper, xl ]
Now if my actual phrase was "what a great super-duper-xl" that would turn into a sequence:
[what,a,great,superduperxl,super,duper,xl]
That matches the phrase "great superduperxl" which is fine.
However, if the phrase was "the super-duper-xl emerged" the sequence would be:
[the,superduperxl,super,duper,xl,emerged]
This does not phrase match "superduperxl emerged", however it would if the part tokens (super,duper,xl) were not emitted.
Is there any way I can concatenate words keeping only the concatenated word and filtering out the word parts?

Answer 1

Pattern replace character filter can be used here.
"-" is replaced with "" to generate tokens
Query
PUT my-index1
{
"settings": {
"analysis": {
"analyzer": {
"remove_hyphen_analyzer": {
"tokenizer": "standard",
"char_filter": [
"remove_hyphen_filter"
]
}
},
"char_filter": {
"remove_hyphen_filter": {
"type": "pattern_replace",
"pattern": "-",
"replacement": ""
}
}
}
},
"mappings": {
"properties": {
"title":{
"type": "text",
"analyzer": "remove_hyphen_analyzer"
}
}
}
}
POST my-index1/_analyze
{
"analyzer": "remove_hyphen_analyzer",
"text": "the super-duper-xl emerged"
}
Result
{
"tokens" : [
{
"token" : "the",
"start_offset" : 0,
"end_offset" : 3,
"type" : "<ALPHANUM>",
"position" : 0
},
{
"token" : "superduperxl",
"start_offset" : 4,
"end_offset" : 18,
"type" : "<ALPHANUM>",
"position" : 1
},
{
"token" : "emerged",
"start_offset" : 19,
"end_offset" : 26,
"type" : "<ALPHANUM>",
"position" : 2
}
]
}