Lowercase normalizer is used for wildcard queries #28894
Description
Elasticsearch version (bin/elasticsearch --version): 6.2.2
Description of the problem including expected versus actual behavior:
Say you index a field Aa as a text field with a Lowercase analyzer.
When you search for aa*, it matches. Searching for Aa* does not match which is normal as the wildcard queries are not analyzed.
Say you index a field Aa as a keyword field with a Lowercase normalizer.
When you search for aa*, it matches. Searching for Aa* matches as well although the wildcard queries are not analyzed.
Steps to reproduce:
DELETE test
PUT test
{
"settings": {
"analysis": {
"normalizer": {
"lowercase_normalizer": {
"type": "custom",
"filter": [
"lowercase"
]
}
}
}
},
"mappings": {
"doc": {
"properties": {
"foo": {
"type": "text",
"analyzer": "simple",
"fields": {
"keyword": {
"type": "keyword",
"normalizer": "lowercase_normalizer"
}
}
}
}
}
}
}
PUT test/doc/1?refresh
{
"foo": "Bbb Aaa"
}
# Does not match -> OK
GET test/_search
{
"query": {
"wildcard": {
"foo": "Bb*"
}
}
}
# Match -> OK
GET test/_search
{
"query": {
"wildcard": {
"foo": "bb*"
}
}
}
# Match but should not -> KO
GET test/_search
{
"query": {
"wildcard": {
"foo.keyword": "Bb*"
}
}
}
# Match -> OK
GET test/_search
{
"query": {
"wildcard": {
"foo.keyword": "bb*"
}
}
}
I spoke with @jpountz who thinks it might be related to https://issues.apache.org/jira/browse/LUCENE-8186
Opening the issue so we can track it.