Configuring a field for an ID match

Elastic Stack Elasticsearch
1

I have a field studyLabel with values like "GSM-ABC123" where "GSM-" is always the same and the second part 'ABC123' is the actual ID. I need a document with such an ID to be matched by ''GSM-ABC123" or "ABC123", but search for "GSM" should not return anything.

This is how I configured the field:

{
  "settings": {
	  "index": {
		  "analysis" : {
			  "filter": {
				  "study_label_id_only_filter" : {
					  "type": "pattern_replace",
					  "pattern": "^GSM-(.*)",
					  "replacement": "$1"
				  }
				},
			  "analyzer": { 
				  "study_label_analyzer": {
					  "tokenizer": "whitespace",
					  "filter": [
						  "study_label_id_only_filter"
					  ]
				  }
			  }
		  }
	  }
	},
	"mappings": {
	  "study": {
		  "properties": {
			  "studyLabel": {
				  "type": "string",
				  "index": "not_analyzed",
				  "fields": {
					  "idOnly": {
						  "type": "string",
						  "analyzer": "study_label_analyzer"
					  }
				  }
			  }
		  }
	  }
  }
}

This way, studyLabel only matches the full id "GSM-ABC123" (and does NOT match "GSM" or "ABC123" ) and studyLabel.idOnly only matches "ABC123" (and does NOT match "GSM" or "GSM-ABC123").

But, when I query match: { "_all", "GSM"} I still get ALL the documents back. Any ideas why?