How to boost documents matching one of the query_string - elasticsearch

Question

Elasticsearch newbie here. I'm trying to lookup documents that has foo in its name but want to prioritize that ones having bar as well i.e. those with bar will be at the top of the list. The result doesn't have the ones with bar at the top. boost here doesn't seem to have any effect, likely I'm not understanding how boost works here. Appreciate any help here.
query: {
bool: {
should: [
{
query_string: {
query: `name:foo*bar*`,
boost: 5
}
},
{
query_string: {
query: `name:*foo*`,
}
}
]
}
}
Sample document structure:
{
"name": "foos, one two three",
"type": "car",
"age": 10
}
{
"name": "foos, one two bar three",
"type": "train",
"age": 30
}
Index mapping
{
"detail": {
"mappings": {
"properties": {
"category": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword",
"ignore_above": 256
}
}
},
"name": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword",
"ignore_above": 256
}
}
},
"servings": {
"properties": {
"name": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword",
"ignore_above": 256
}
}
}
}
}
}
}
}
}

Answer 1

Try switching the order for the query like so:
query: {
bool: {
should: [
{
query_string: {
query: `name:*foo*`,
}
},
{
query_string: {
query: `name:foo*bar*`,
boost: 5
}
}
]
}
}
it should work but if not you might need to do a nested search.

Answer 2

Search against keyword field.
If you will only run first part of the query ("query": "name:foo*bar*"), you will see that it is not returning anything. It is searching against tokens generated rather than whole string.
Text "foos, one two bar three" generates tokens like ["foos","one","two","bar","three"] and query is searching for "foo*bar*" in individual tokens hence no result. Keyword fields are stored as it is so search is happening against entire text.
{
"query": {
"bool": {
"should": [
{
"query_string": {
"query": "name.keyword:foo*bar*",
"boost": 5
}
},
{
"query_string": {
"query": "name.keyword:*foo*"
}
}
]
}
}
Wildcards take huge memory and don't scale well. So it is better to avoid it. If foo and bar appear at start of words , you can use prefix query
{
"query": {
"bool": {
"should": [
{
"prefix": {
"name": "foo"
}
},
{
"prefix": {
"name": "bar"
}
}
]
}
}
}
You can also explore ngrams