Upgrading from Elasticsearch 5.x to 6.x. We make extensive use of query string queries and commonly construct queries which used fields of different types.
In 5.x, the following query worked correctly and without error:
{
"query": {
"query_string": {
"query": "my_keyword_field:\"Exact Phrase Here\" my_text_field:(any words) my_other_text_field:\"Another phrase here\" date_field:[2018-01-01 TO 2018-05-01]",
"default_operator": "AND",
"analyzer": "custom_text"
}
}
}
In 6.x, this query will return the following error:
{
"type": "illegal_state_exception",
"reason": "field:[my_keyword_field] was indexed without position data; cannot run PhraseQuery"
}
If I wrap the phrase in parentheses instead of quotes, the search will return 0 results:
{
"query": {
"query_string": {
"query": "my_keyword_field:(Exact Phrase Here)",
"default_operator": "AND",
"analyzer": "custom_text"
}
}
}
I guess this is because there is a conflict between the way the analyzer stems the incoming query and how the data is stored in the keyword field, but the phrase approach (my_keyword_field:"Exact Phrase Here") did work in 5.x.
Is this no longer supported in 6.x? And if not, what is the migration path and/or a good workaround?
It would be better to rephrase the query by using different type of queries available for different use cases. For example use term query for exact search on keyword field. Use range query for ranges etc.
You can rephrase query as below:
{
"query": {
"bool": {
"must": [
{
"query_string": {
"query": "my_text_field:(any words) my_other_text_field:\"Another phrase here\"",
"default_operator": "AND",
"analyzer": "custom_text"
}
},
{
"term": {
"my_keyword_field": "Exact Phrase Here"
}
},
{
"range": {
"date_field": {
"gte": "2018-01-01",
"lte": "2018-05-01"
}
}
}
]
}
}
}
Related
I'm using query string to search on documents in my index.
GET my_index/_search
{
"query": {
"bool": {
"must": [{
"query_string": {
"query": "table test",
"default_field": "table.name",
"default_operator":"AND"
}
}]
}
}
}
the problem is that it returns all additional strings that include search keywords.. I wanna to give strings that have exact phrase.
for example the documents table test 1 and table test 12 and table test are in my index. when I search table test, I wanna it just return table test.
I used term also, but it could not consider space charter between strings!
how can I handle this?
your mapping is generated by Elasticsearch, than for every text field there will be a corresponding .keyword field and hence
{
"query": {
"term": {
"table.name.kwyword": { // Note .keyword in the field name.
"value": "table test",
"boost": 1.0
}
}
}
if you don't have a .keyword field, then you have to create a keyword field and use term query that is used for exact or keyword searches.
You can use Match Phrase Query as Amit suggested in another answer.
Also, if you want to use only Query String type of query then you can give your query in double quotes as shown below:
GET my_index/_search
{
"query": {
"bool": {
"must": [{
"query_string": {
"query": "\"table test\"",
"default_field": "table.name",
"default_operator":"AND"
}
}]
}
}
}
Updated:
if you want to do exact match in entire field then you can go ahead with term query in elasticsearch:
{
"query": {
"term": {
"table.name.keyword": {
"value": "table test",
"boost": 1.0
}
}
}
}
My knowledge of Elasticsearch is a bit limited, so what I want to do might not even be possible.
Say I have an ecommerce where I want to be able to freely search on the article names and other fields, but I also want to search on exact article codes aswell. Is this possible in the same query?
Example:
"articlecode": "v400",
"name": "Earplugs for humans"
}
{
"articlecode": "b6655",
"name": "Hammer 400"
}
So can a query be written that combines both multimatch and terms? So that If I search for '400' I get 2 results, but if I search for v400 I just get one result as it is an exact match on the "articlecode"-field.
Below is our current query, where i have an ngram on the "name" field and where I use the term-keyword on the language-field.
{
"size": 10,
"query": {
"bool": {
"must": {
"multi_match": {
"query": "v400",
"fields": [
"articlecode^10",
"name^7"
]
}
},
"filter": {
"term": {
"IdLang.keyword": "sv"
}
}
}
}
}
Have you ever thought of using query_string instead of multi_match? Then you can use wildcard in your search:
{
"size": 10,
"query": {
"bool": {
"must": {
"query_string": {
"query": "*v400",
"fields": [
"articlecode^10",
"name^7"
]
}
}
}
}
}
If you want to search with 400 anywhere in the 2 fields, you can do *400*, or only leading or trailing, depending on what you want.
Works as expected:
{
"query": {
"query_string": {
"query": "Hofstetten-Grünau"
}
}
}
an added wildcard at the end delivers no results and I wonder why:
{
"query": {
"query_string": {
"query": "Hofstetten-Grünau*"
}
}
}
how to fix it?
elasticsearch v5.3.2
This delivers results:
{
"query": {
"query_string": {
"query": "Hofstetten*"
}
}
}
I use a single search field. The end user can freely use wildcards as they see fit. A user might type in:
hofstetten grünau
+ort:hofstetten-grünau
+ort:Hofstetten-G*
so using a match query wont work out for me.
I am using Jest (Java Annotations) as Mapping, and using "default" for this field. My index mapping declares nothing special for the field:
{
"mappings": {
"_default_": {
"date_detection": false,
"dynamic_templates": [{
}]
}
}
}
Adding the wildcard "*" at the end of your query string is causing the query analyzer to interpret the dash between "Hofstetten" and "Grünau" as a logical NOT operator. So you're actually searching for documents that contain Hofstetten but do NOT contain Grünau.
You can verify this by doing the following variations of your search:
"query": "Hofstetten-XXXXX" #should not return results
"query": "Hofstetten-XXXXX*" #should return results
To fix this I would recommend using a match query instead of a query_string query:
{"query": {"match": { "city": "Hofstetten-Grünau" }}}'
(with whatever your appropriate field name is in place of city).
I am working to match a 'term' to multi fields (or _all field)
I want to do a fuzzy match on cross_fields but it is not supported.
any ideas how to do it or any other ways to do it ?
query: {
multi_match: {
query: term,
type: "cross_fields",
fields: ['_all']
}
}
when trying the solution here
ElasticSearch multi_match query over multiple fields with Fuzziness
I get this error
[parsing_exception] Fuziness not allowed for type [cross_fields], with
{ line=1 & col=128 }
elasticsearch version 5.0
edit:
here is the query I am building
bool: {
must: [
{
fuzzy: {
_all: term
}
},
{
fuzzy: {
"location.country": country
}
},
{
fuzzy: {
"location.city": city
}
}
]
}
cross_fields works by searching the term on your multiple fields. Since fuzziness isn't supported for cross_fields you have to write the query in a different way.
One possible is: implement your own "cross_fields" with shoulds and add there the fuzziness.
Say your term is: "term1 term2", you can split by word boundary (Regex \b) then should them in this form:
{
{
"query": {
"bool": {
"should": [{
"match": {
"field1": "term",
"fuzziness": 1
}
},{
"match": {
"field1": "term",
"fuzziness": 1
}
},{
"match": {
"field2": "term1",
"fuzziness": 1
}
},{
"match": {
"field2": "term12",
"fuzziness": 1
}
}
]
}
}
}
}
This is probably less the optimal if you have many fields, the query will become a cartesian product of the terms and fields.
Important note You're using _all field which is one field. which all other fields are indexed into. Maybe you don't even need cross_fields?
Perhaps it's a basic question; by the way, I need to search in every indexed field and to have a specific fixed value for another field.
How can I do it?
Currently I have a simple: query( "aValue", array_of_models )
I tried many options without success, for example:
query({
"query": {
"bool": {
"query": "aValue",
"filter": {
"term": {
"published": "true"
}
}
}
}
})
I would prefer to avoid to specify the fields to search in because I use the same search params for different models.
I found a solution, perhaps it's not optimized but works:
{
"query": {
"bool": {
"should": [
{
"match": {
"_all": "aValue"
}
}
],
"filter": {
"term": {
"published": true
}
}
}
}
}
Not sure if I understood correctly your intention.
The _all field is as default enabled. So if you have no special mapping every indexed field value is added as text string to the _all field.
You can use the
Query String Query, https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-query-string-query.html
Simple Query String Query, https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-simple-query-string-query.html
With a simple query like this, that should work for you.
GET my_index/_search
{
"query": {
"simple_query_string": {
"query": "aValue",
"fields": []
}
}
}
Both query types contains parameters, that should suffice your use case IMHO.