Why the field is not subject to lexical analysis - azure

Question

I'm trying to make my searches ignore word accents
To do this I decided to use the language analyzer: es.microsoft
I was testing the analyzer with the word "Lámpara" in the analyzer API and I got the following results:
{
"token": "lampara",
"startOffset": 0,
"endOffset": 7,
"position": 0
},
{
"token": "lámpara",
"startOffset": 0,
"endOffset": 7,
"position": 0
}
I have only 2 documents in my test index:
{
"#search.score": 1,
"Id": "2",
"Nombre": "Lampara"
},
{
"#search.score": 1,
"Id": "1",
"Nombre": "Lámpara"
}
When searching for the word in the index search=Lámpara I get the following results:
{
"#search.score": 0.30685282,
"Id": "1",
"Nombre": "Lámpara"
}
For what reason the document is only received with Nombre = "Lámpara" and not Nombre = "Lampara" (without accent). I have the impression that the Name field was not sent to the lexical analysis
The definition of my index is as follows
{
"name": "test",
"fields": [
{
"name": "Id",
"type": "Edm.String",
"facetable": false,
"filterable": true,
"key": true,
"retrievable": true,
"searchable": false,
"sortable": false,
"analyzer": null,
"indexAnalyzer": null,
"searchAnalyzer": null,
"synonymMaps": [],
"fields": []
},
{
"name": "Nombre",
"type": "Edm.String",
"facetable": false,
"filterable": false,
"key": false,
"retrievable": true,
"searchable": true,
"sortable": false,
"analyzer": "es.microsoft",
"indexAnalyzer": null,
"searchAnalyzer": null,
"synonymMaps": [],
"fields": []
}
],
"suggesters": [],
"scoringProfiles": [],
"defaultScoringProfile": null,
"corsOptions": null,
"analyzers": [],
"charFilters": [],
"tokenFilters": [],
"tokenizers": []
}
I would appreciate any help, and an apology for my bad English

Answer 1

Sorry for the delay in getting your an answer. Indeed, the Microsoft Spanish analyzer currently only fold the accents in documents, so they can be matched by queries that forgo the accents (as you mentioned, search for Lampara, it will match documents that contains Lámpara, but if you explicitly set the accents in the query (for example, searching for Lámpara), it won't match documents that don't have any accents.
If this behavior is important to you, you can instead use the es.lucene analyzer which actually actually do "ascii folding" (removing of accents) both at indexing and at search time.