Mapping not analyzed field, but making it not case sensitive

Elastic Stack Elasticsearch
1

Hello, I'm trying to migrate from Solr, but I'm having trouble
understanding mapping in ES.
I need to have just a plain string field for things like user name or
category name. But it could be not case sensetive.
So if I have a field with value like "userName":"Eugene Strokin" the
document should be returned in search like this: "userName":"eugene
strokin", as well as "userName":"EUGENE Strokin", and
"userName":"EugEne StrOKin"
I've put a mapping file with field description:
"userName":{"type":"string", index:"not_analyzed"},
But of course it is case sensitive.
So, query like this will not return anything:
{
"filtered" : {
"query" : {
"match_all" : {
}
},
"filter" : {
"term" : {
"userName" : "tessa"
}
}
}
}
if user name spelled using wrong case.
Could you please let me know how to mapped not case sensitive strings?

Thank you,
Eugene S.

2

You can configure an analyzer with keyword tokenizer and lowercase filter,
call it "keyword_lowercase" or something similar, and then, map the
relevant fields with this analyzer.

On Fri, Nov 25, 2011 at 11:42 PM, Eugene Strokin eugene@strokin.infowrote:

Hello, I'm trying to migrate from Solr, but I'm having trouble
understanding mapping in ES.
I need to have just a plain string field for things like user name or
category name. But it could be not case sensetive.
So if I have a field with value like "userName":"Eugene Strokin" the
document should be returned in search like this: "userName":"eugene
strokin", as well as "userName":"EUGENE Strokin", and
"userName":"EugEne StrOKin"
I've put a mapping file with field description:
"userName":{"type":"string", index:"not_analyzed"},
But of course it is case sensitive.
So, query like this will not return anything:
{
"filtered" : {
"query" : {
"match_all" : {
}
},
"filter" : {
"term" : {
"userName" : "tessa"
}
}
}
}
if user name spelled using wrong case.
Could you please let me know how to mapped not case sensitive strings?

Thank you,
Eugene S.

3

Is anyway to define this new type on some configuration file?
I'm trying to put:
index :
analysis:
analyzer:
string_lowercase:
tokenizer : keyword
filter : lowercase

in config/elasticsearch.yml file which is under config folder, but I
keep getting
Caused by: org.elasticsearch.index.mapper.MapperParsingException: No
handler for type [string_lowercase] declared on field [screenName]

Thank you,
E.S.

On Nov 28, 6:58 am, Shay Banon kim...@gmail.com wrote:

You can configure an analyzer with keyword tokenizer and lowercase filter,
call it "keyword_lowercase" or something similar, and then, map the
relevant fields with this analyzer.

On Fri, Nov 25, 2011 at 11:42 PM, Eugene Strokin eug...@strokin.infowrote:

Hello, I'm trying to migrate from Solr, but I'm having trouble
understanding mapping in ES.
I need to have just a plain string field for things like user name or
category name. But it could be not case sensetive.
So if I have a field with value like "userName":"Eugene Strokin" the
document should be returned in search like this: "userName":"eugene
strokin", as well as "userName":"EUGENE Strokin", and
"userName":"EugEne StrOKin"
I've put a mapping file with field description:
"userName":{"type":"string", index:"not_analyzed"},
But of course it is case sensitive.
So, query like this will not return anything:
{
"filtered" : {
"query" : {
"match_all" : {
}
},
"filter" : {
"term" : {
"userName" : "tessa"
}
}
}
}
if user name spelled using wrong case.
Could you please let me know how to mapped not case sensitive strings?

Thank you,
Eugene S.

4

Seems to work: config · GitHub. Are you sure you configured
the yml correctly?

On Tue, Nov 29, 2011 at 2:40 AM, Eugene Strokin eugene@strokin.info wrote:

Is anyway to define this new type on some configuration file?
I'm trying to put:
index :
analysis:
analyzer:
string_lowercase:
tokenizer : keyword
filter : lowercase

in config/elasticsearch.yml file which is under config folder, but I
keep getting
Caused by: org.elasticsearch.index.mapper.MapperParsingException: No
handler for type [string_lowercase] declared on field [screenName]

Thank you,
E.S.

On Nov 28, 6:58 am, Shay Banon kim...@gmail.com wrote:

You can configure an analyzer with keyword tokenizer and lowercase
filter,
call it "keyword_lowercase" or something similar, and then, map the
relevant fields with this analyzer.

On Fri, Nov 25, 2011 at 11:42 PM, Eugene Strokin <eug...@strokin.info
wrote:

Hello, I'm trying to migrate from Solr, but I'm having trouble
understanding mapping in ES.
I need to have just a plain string field for things like user name or
category name. But it could be not case sensetive.
So if I have a field with value like "userName":"Eugene Strokin" the
document should be returned in search like this: "userName":"eugene
strokin", as well as "userName":"EUGENE Strokin", and
"userName":"EugEne StrOKin"
I've put a mapping file with field description:
"userName":{"type":"string", index:"not_analyzed"},
But of course it is case sensitive.
So, query like this will not return anything:
{
"filtered" : {
"query" : {
"match_all" : {
}
},
"filter" : {
"term" : {
"userName" : "tessa"
}
}
}
}
if user name spelled using wrong case.
Could you please let me know how to mapped not case sensitive strings?

Thank you,
Eugene S.

5

HI Kimchy,
I tried your gist and it gave me following exception

{"error":"RemoteTransportException[[Banner, Robert
Bruce][inet[/192.168.1.15:9300]][indices/create]]; nested:
MapperParsingException[mapping [type1]]; nested:
MapperParsingException[Analyzer [string_lowercase] not found for field
[field1]];

I first created the analyzer in the elasticsearch.yml ,as provided in your
gist, and after that I created the mapping. Please let me know what I am
doing wrong.

On Tuesday, 29 November 2011 13:15:40 UTC+5:30, kimchy wrote:

Seems to work: config · GitHub. Are you sure you
configured the yml correctly?

On Tue, Nov 29, 2011 at 2:40 AM, Eugene Strokin eugene@strokin.infowrote:

Is anyway to define this new type on some configuration file?
I'm trying to put:
index :
analysis:
analyzer:
string_lowercase:
tokenizer : keyword
filter : lowercase

in config/elasticsearch.yml file which is under config folder, but I
keep getting
Caused by: org.elasticsearch.index.mapper.MapperParsingException: No
handler for type [string_lowercase] declared on field [screenName]

Thank you,
E.S.

On Nov 28, 6:58 am, Shay Banon kim...@gmail.com wrote:

You can configure an analyzer with keyword tokenizer and lowercase
filter,
call it "keyword_lowercase" or something similar, and then, map the
relevant fields with this analyzer.

On Fri, Nov 25, 2011 at 11:42 PM, Eugene Strokin <eug...@strokin.info
wrote:

Hello, I'm trying to migrate from Solr, but I'm having trouble
understanding mapping in ES.
I need to have just a plain string field for things like user name or
category name. But it could be not case sensetive.
So if I have a field with value like "userName":"Eugene Strokin" the
document should be returned in search like this: "userName":"eugene
strokin", as well as "userName":"EUGENE Strokin", and
"userName":"EugEne StrOKin"
I've put a mapping file with field description:
"userName":{"type":"string", index:"not_analyzed"},
But of course it is case sensitive.
So, query like this will not return anything:
{
"filtered" : {
"query" : {
"match_all" : {
}
},
"filter" : {
"term" : {
"userName" : "tessa"
}
}
}
}
if user name spelled using wrong case.
Could you please let me know how to mapped not case sensitive strings?

Thank you,
Eugene S.