Elasticsearch Alias第二篇 Alias 过滤器和路由

主要针对1.5版本，主要参考自官网资料，可以理解为一个翻译+实践+扩充的版本

Filtered Alias

带过滤器的别名提供了一个十分简单的方式来创建同一个index下的不同的“展示”。过滤器能够使用Query DSL来定义并且被应用到所有的搜索，计数，通过查询删除和其它类似的行为。

为了创建一个带过滤器的别名，首先需要确保所有的字段都存在于mapping中。

创建一个mapping

curl -XPUT 'http://localhost:9200/test2' -d '{
  "mappings": {
    "type1": {
      "properties": {
        "user" : {
          "type": "string",
          "index": "not_analyzed"
        }
      }
    }
  }
}'

我们可以创建一个在user字段带过滤器的别名。

curl -XPOST 'http://localhost:9200/_aliases' -d '{
    "actions" : [
        {
            "add" : {
                 "index" : "test2",
                 "alias" : "alias10",
                 "filter" : { "term" : { "user" : "kimchy" } }
            }
        }
    ]
}'

成功则返回

{"acknowledged":true}

我们向test2中间放入两个user，trump和kimchy，如下

~ % curl -XPOST 'localhost:9200/alias10/testtype' -d '{"user":"trump"}'
{"_index":"test2","_type":"testtype","_id":"AVhnwgnAu0MIai42Pwvm","_version":1,"created":true}% 
~ % curl -XPOST 'localhost:9200/alias10/testtype' -d '{"user":"kimchy"}'
{"_index":"test2","_type":"testtype","_id":"AVhnwkRiu0MIai42Pwvn","_version":1,"created":true}%

可以看到，我们通过别名进行文档输入，虽然alias10为带过滤器，限制了user为kimchy，但是我们通过alias10的别名进行user非kimchy文件的输入还是成功。进行搜索，得到结果如下：

~ % curl 'localhost:9200/test2/_search' -d '{"query":{"match_all":{}}}'
{"took":2,"timed_out":false,"_shards":{"total":5,"successful":5,"failed":0},"hits":{"total":2,"max_score":1.0,"hits":[{"_index":"test2","_type":"testtype","_id":"AVhnwkRiu0MIai42Pwvn","_score":1.0,"_source":{"user":"kimchy"}},{"_index":"test2","_type":"testtype","_id":"AVhnwgnAu0MIai42Pwvm","_score":1.0,"_source":{"user":"trump"}}]}}% 
~ % curl 'localhost:9200/alias10/_search' -d '{"query":{"match_all":{}}}'
{"took":17,"timed_out":false,"_shards":{"total":5,"successful":5,"failed":0},"hits":{"total":1,"max_score":1.0,"hits":[{"_index":"test2","_type":"testtype","_id":"AVhnwkRiu0MIai42Pwvn","_score":1.0,"_source":{"user":"kimchy"}}]}}%

可以看到，通过test2这个index直接进行搜索可以看到之前放入的两个文档，但是通过alias10这个别名就只能看到符合过滤器过滤后的结果了，只有一个user为kimchy的结果。

Routing路由

能够为别名绑上路由。这个特性可以和过滤器的特性一起使用，来避免不必要地分片操作。

如下的命令创建了一个新的别名alias2指向了index test。在alias2被创建之后，所有和这个别名相关的操作将会被自动修正为使用值为1的路由。

curl -XPOST 'http://localhost:9200/_aliases' -d '
{
    "actions" : [
        {
            "add" : {
                 "index" : "test",
                 "alias" : "alias2",
                 "routing" : "1"
            }
        }
    ]
}'

也可以分开指定搜索和索引的路由

curl -XPOST 'http://localhost:9200/_aliases' -d '
{
    "actions" : [
        {
            "add" : {
                 "index" : "test",
                 "alias" : "alias2",
                 "search_routing" : "1,2",
                 "index_routing" : "2"
            }
        }
    ]
}'

进行测试，同时使用路由和过滤器

curl -XPOST 'http://localhost:9200/_aliases' -d '
{
    "actions" : [
        {
            "add" : {
                 "index" : "test",
                 "alias" : "alias2",
                 "search_routing" : "2,3",
                 "index_routing" : "1",
                 "filter": {"term": {"user": "kitty"}}
            }
        }
    ]
}'

注意啦，index_routing不能指向多个值，不然会报错

{"error":"ElasticsearchIllegalArgumentException[alias [alias2] has several index routing values associated with it]","status":400}%

再加一个alias3

curl -XPOST 'http://localhost:9200/_aliases' -d '
{
    "actions" : [
        {
            "add" : {
                 "index" : "test",
                 "alias" : "alias3",
                 "search_routing" : "1",
                 "index_routing" : "2",
                 "filter": {"term": {"user": "hello"}}
            }
        }
    ]
}'

测试如下：

~ % curl -XPOST 'localhost:9200/alias2/testtest' -d '{"user": "trump"}'
{"_index":"test","_type":"testtest","_id":"AVhqrXnwYhUIwOtlirMh","_version":1,"created":true}% 
~ % curl -XPOST 'localhost:9200/alias2/testtest' -d '{"user": "hello"}'
{"_index":"test","_type":"testtest","_id":"AVhqrcWRYhUIwOtlirMi","_version":1,"created":true}% 
~ % curl -XPOST 'localhost:9200/alias2/testtest' -d '{"user": "world"}'
{"_index":"test","_type":"testtest","_id":"AVhqrdp7YhUIwOtlirMj","_version":1,"created":true}% 
~ % curl -XPOST 'localhost:9200/alias3/testwhat' -d '{"user": "kitty"}'
{"_index":"test","_type":"testwhat","_id":"AVhqrhRyYhUIwOtlirMk","_version":1,"created":true}% 
~ % curl -XPOST 'localhost:9200/alias3/testwhat' -d '{"user": "omg"}'
{"_index":"test","_type":"testwhat","_id":"AVhqriccYhUIwOtlirMl","_version":1,"created":true}%                                                                                                                                                                                ~ % curl -XPOST 'localhost:9200/alias3/testwhat' -d '{"user": "haha"}'
{"_index":"test","_type":"testwhat","_id":"AVhqrjmCYhUIwOtlirMm","_version":1,"created":true}% 
~ % curl -XPOST 'localhost:9200/alias2/_search'
{"took":29,"timed_out":false,"_shards":{"total":2,"successful":2,"failed":0},"hits":{"total":1,"max_score":1.0,"hits":[{"_index":"test","_type":"testwhat","_id":"AVhqrhRyYhUIwOtlirMk","_score":1.0,"_source":{"user": "kitty"}}]}}%
~ % curl -XPOST 'localhost:9200/alias3/_search'
{"took":3,"timed_out":false,"_shards":{"total":1,"successful":1,"failed":0},"hits":{"total":1,"max_score":1.0,"hits":[{"_index":"test","_type":"testtest","_id":"AVhqrcWRYhUIwOtlirMi","_score":1.0,"_source":{"user": "hello"}}]}}%
~ % curl -XDELETE 'localhost:9200/alias3/testwhat/AVhqrhRyYhUIwOtlirMk'
{"found":true,"_index":"test","_type":"testwhat","_id":"AVhqrhRyYhUIwOtlirMk","_version":2}% 
~ % curl -XPOST 'localhost:9200/alias2/testwhat' -d '{"user": "kitty"}'
{"_index":"test","_type":"testwhat","_id":"AVhqtv_HYhUIwOtlirMn","_version":1,"created":true}%
~ % curl -XPOST 'localhost:9200/alias2/_search'
{"took":1,"timed_out":false,"_shards":{"total":2,"successful":2,"failed":0},"hits":{"total":0,"max_score":null,"hits":[]}}%

这个测试例子比较有意思，alias2 index的路由为1，搜索的路由为2，3，alias3 index的路由为2，搜索的路由为1，所以使用alias2增加的文档能够被使用alias3搜索到，而使用alias2搜索不到，使用alias3增加的文档能够被alias2搜索到，而使用alias3搜索不到。再加上过滤器，所以alis2搜索match_all只能够得到kitty，alias3搜索match_all只能够得到hello，如果我们通过alias3删除了kitty，通过alias2增加了kitty，可以看到，alias2也是搜索不到的。