Searching with multiple values - JAVA

issam · March 9, 2016, 7:03pm

Hello,

I have to search on the indexes already created with multiple values in a given field, but I always get nothing in the result

Can you tell please if it is ok with the following code (I use elasticsearch Java API)

query = QueryBuilders.termsQuery(param, values);
boolQueryBuilder.must(query);

I have something like that in boolQueryBuilder

{
"bool" : {
```
"must" : [ {
```
```
  "terms" : {
```

    "param1" : [ value1, value2, value3 ]

```
  }
```
```
}, {
```
```
  "wildcard" : {
```
```
    "param2" : "*"
```
```
  }
```
```
} ]
```

I'm sure, that value1, 2 & 3 already exist, but I always get nothing, is there anyone who has an idea please ?

Thank you

issam · March 9, 2016, 7:23pm

well, that was quickly, it is ok with using the termsQuery search with many values in one a single field, the problem here, is that my values contain "-" wich is a special character, so elasticsearch replace it with space (I think)

If any one has an answer, for this problem (special characters) I'll be really thankful

cbuescher · March 9, 2016, 9:00pm

Take a look at the different Tokenizers, the default is the Standard tokenizer which splits on some characters (as you have seen). Maybe start with the Whitespace tokenizer instead.

issam · March 10, 2016, 5:16pm

Thank you for Christoph, I found how to modify the tokenizer to whistespace one (using JAVA) but now, the wildcard terms are not working anymore, '*' and '?' are not considered token chars anymore, which is not good because I need these characters in my search

Is there anyway to keep '*' and '?' and remove only other characters such as '-' and '@' from token chars ??

Thank you

cbuescher · March 11, 2016, 2:18pm

I'm afraid I'm not sure I understand what you are trying to do here. This wildcard query basically matches everything in the param2 field. Also I don't quiet understand what you mean by "'*' and '?' are not considered token chars anymore". Maybe you can give an example of a document you are trying to find and which isn't returned by the query you are using?