Elasticsearch: Browse sorted terms - elasticsearch

Question

Let's say i have a set of document with common personal information like:
{
"name": "Samuel",
"Job": "Engineer",
....
}
My customer want to browse through all the terms in the "name" field, starting from a page centred on a "selected" name.
For example they want to search for "Samuel" and show a page of 7 elements like:
Eddie: 7
Lucian: 3
Sammy: 1
Samuel: 3
Scott:3
Teddy: 2
Tod: 1
Where the names are sorted alphabetically and the number are the number of occurrence.
Also would be nice to be able to go up and down through pages.
This is just an example, in the reality I may have lot and lot of unique keys to browse, so returning all the terms and loop over them is not really a solution.
There is a way to achieve this with ElasticSearch ?

Answer 1

I'd suggest using filtered aggs based on bucketing the users by the name first char -- so from a to g, h to n etc:
Syncing a few docs:
POST jobs/_doc
{"name":"Samuel","Job":"Engineer"}
POST jobs/_doc
{"name":"Lucian","Job":"Engineer"}
POST jobs/_doc
{"name":"Teddy","Job":"Engineer"}
POST jobs/_doc
{"name":"Tod","Job":"Engineer"}
POST jobs/_doc
{"name":"Andrew","Job":"Engineer"}
Then applying scripted filters:
GET jobs/_search
{
"size": 0,
"aggs": {
"alpha_buckets": {
"filter": {
"match_all": {}
},
"aggs": {
"a_g": {
"filter": {
"script": {
"script": {
"source": """
def first_char = doc['name.keyword'].value.toLowerCase().toCharArray()[0];
def char_int = Character.getNumericValue(first_char);
return char_int >= 10 && char_int <= 16
""",
"lang": "painless"
}
}
},
"aggs": {
"a_g": {
"terms": {
"field": "name.keyword",
"order": {
"_term": "asc"
}
}
}
}
},
"h_n": {
"filter": {
"script": {
"script": {
"source": """
def first_char = doc['name.keyword'].value.toLowerCase().toCharArray()[0];
def char_int = Character.getNumericValue(first_char);
return char_int >= 17 && char_int <= 23
""",
"lang": "painless"
}
}
},
"aggs": {
"h_n": {
"terms": {
"field": "name.keyword",
"order": {
"_term": "asc"
}
}
}
}
}
}
}
}
}
yielding
"aggregations" : {
"alpha_buckets" : {
"doc_count" : 7,
"h_n" : {
"doc_count" : 2,
"h_n" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "Lucian",
"doc_count" : 2
},
{
"key" : "Lupe",
"doc_count" : 1
}
]
}
},
"a_g" : {
"doc_count" : 1,
"a_g" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "Andrew",
"doc_count" : 1
},
{
"key" : "Anton",
"doc_count" : 1
}
]
}
}
}
Now, you don't need to get all alphabetical buckets at once, the above is just to show what's possible.