Elasticsearch NEST Library - elasticsearch

Question

We need to create facets in a field which is an array of int [1,2,3,4,5,6].
If we map this field as Text, Elasticsearch throws an exception. If we mark it as Keyword we don't get all the individual items in the facet.
How can this be achieved with the NEST library (version 5)
Many thanks

Answer 1

Map it as an integer. NEST's automapping will infer an integer mapping for it by default
Given the following type
public class Message
{
public int[] Integers { get; set; }
}
Automapping would look like
var client = new ElasticClient();
client.CreateIndex("messages", c => c
.Mappings(m => m
.Map<Message>(mm => mm
.AutoMap()
)
)
);
which produces the following mapping
{
"mappings": {
"message": {
"properties": {
"integers": {
"type": "integer"
}
}
}
}
}
You can use attribute mapping if you want to explicitly control the mapping
public class Message
{
[Number(NumberType.Integer)]
public int[] Integers { get; set; }
}
var client = new ElasticClient();
client.CreateIndex("messages", c => c
.Mappings(m => m
.Map<Message>(mm => mm
.AutoMap()
)
)
);
Or fluent mapping if you prefer to keep mapping separate from your POCOs
client.CreateIndex(messagesIndex, c => c
.Mappings(m => m
.Map<Message>(mm => mm
.AutoMap()
.Properties(p => p
.Number(n => n
.Name(nn => nn.Integers)
.Type(NumberType.Integer)
)
)
)
)
);
or using the scalar method
client.CreateIndex(messagesIndex, c => c
.Mappings(m => m
.Map<Message>(mm => mm
.AutoMap()
.Properties(p => p
.Scalar(n => n.Integers)
)
)
)
);
With the mapping in place, faceting on integer values would look like
client.IndexMany(new[] {
new Message { Integers = new [] { 1, 2, 3 } },
new Message { Integers = new [] { 3, 4, 5 } },
new Message { Integers = new [] { 1, 4, 5 } },
});
client.Refresh(messagesIndex);
var searchResponse = client.Search<Message>(s => s
.Size(0)
.Aggregations(a => a
.Terms("ints", t => t
.Field(f => f.Integers)
)
)
);
foreach (var bucket in searchResponse.Aggs.Terms<int>("ints").Buckets)
{
Console.WriteLine($"{bucket.Key}:{bucket.DocCount}");
}
which yields
1:2
3:2
4:2
5:2
2:1