how to ignore number_format_exception error in elasticsearch query - elasticsearch

Question

Hii how to ignore datatype error in below query since it throws error when a string value is provided for a field that has non-numeric(long) datatype. I am aware of lenient parameter but it does not work with term query.
GET employee/_search
{
"query": {
"bool": {
"must": [
{
"bool": {
"should": [
{
"terms": {
"employee_id": [
"abcdef"
]
}
},
{
"terms": {
"employee_name": [
"abcdef"
]
}
}
]
}
}
]
}
}
}
Error Message
"caused_by": {
"type": "number_format_exception",
"reason": "For input string: \"abcdef\""
}
Elasticsearch details
"version" : {
"number" : "7.1.1",
"build_flavor" : "oss",
"build_type" : "tar",
"build_hash" : "Unknown",
"build_date" : "2020-11-03T08:48:42.499923Z",
"build_snapshot" : false,
"lucene_version" : "8.0.0",
"minimum_wire_compatibility_version" : "6.8.0",
"minimum_index_compatibility_version" : "6.0.0-beta1"
}
So the expected behaviour is the datatype error should be ignored and the rest of query runs and produces the result since it is in a should condition and if there is a must condition then give no result
mapping of index
{
"employee" : {
"mappings" : {
"dynamic" : "true",
"properties" : {
"employee_id" : {
"type" : "long"
}
}
}
}
}

Answer 1

You can use the query_string_query instead:
{
"query": {
"bool": {
"must": [
{
"bool": {
"should": [
{
"query_string": {
"query": "employee_id:abcdef" <---
}
},
{
"terms": {
"employee_name": [
"abcdef"
]
}
}
]
}
}
]
}
}
Your original query was a terms query which is equivalent to a logical OR. As such, you can adapt the query string to be:
"employee_id:(abcdef OR xyz OR 123)"
where the value type won't play a role.