Elasticsearch核心知识篇(26)
上机动手实战演练mget批量查询api
批量查询的好处
就是一条一条的查询,比如说要查询100条数据,那么就要发送100次网络请求,这个开销还是很大的 如果进行批量查询的话,查询100条数据,就只要发送1次网络请求,网络请求的性能开销缩减100倍
mget的语法
- 一条一条的查询
GET /test_index/test_type/1
GET /test_index/test_type/2
- mget批量查询
GET /_mget
{
"docs":[
{
"_index":"waws_index",
"_type":"waws_type",
"_id":1
},{
"_index":"waws_index",
"_type":"waws_type",
"_id":2
}
]
}
{
"docs": [
{
"_index": "waws_index",
"_type": "waws_type",
"_id": "1",
"_version": 1,
"found": true,
"_source": {
"name": "waws1"
}
},
{
"_index": "waws_index",
"_type": "waws_type",
"_id": "2",
"_version": 1,
"found": true,
"_source": {
"name": "waws2"
}
}
]
}
- 如果查询的document是一个index下的不同type种的话
GET /waws_index/_mget
{
"docs":[
{
"_type":"waws_type",
"_id":1
},{
"_type":"waws_type",
"_id":2
}
]
}
- 如果查询的数据都在同一个index下的同一个type下,最简单了
GET /waws_index/waws_type/_mget
{
"ids":[1,2]
}
mget的重要性
可以说mget是很重要的,一般来说,在进行查询的时候,如果一次性要查询多条数据的话,那么一定要用batch批量操作的api 尽可能减少网络开销次数,可能可以将性能提升数倍,甚至数十倍,非常非常之重要
Elasticsearch核心知识篇(27)
分布式文档系统_上机动手实战演练bulk批量增删改
bulk语法
POST /_bulk
{ "delete": { "_index": "waws_index", "_type": "waws_type", "_id": "1" }}
{ "create": { "_index": "waws_index", "_type": "waws_type", "_id": "3" }}
{ "waws_name":"waws3" }
{ "index": { "_index": "waws_index", "_type": "waws_type", "_id": "4" }}
{ "waws_name": "waws4" }
{ "update": { "_index": "waws_index", "_type": "waws_type", "_id": "3", "_retry_on_conflict" : 3} }
{ "doc" : {"waws_name" : "update_waws_name"} }
每一个操作
要两个json串
,语法如下:
{"action": {"metadata"}}
{"data"}
举例,比如你现在要创建一个文档,放bulk里面,看起来会是这样子的:
{"index": {"_index": "waws_index", "_type", "waws_type", "_id": "1"}}
{"waws_field1": "waws1", "waws_field2": "waws2"}
有哪些类型的操作可以执行呢?
- delete:删除一个文档,只要1个json串就可以了
- create:PUT /index/type/id/_create,强制创建
- index:普通的put操作,可以是创建文档,也可以是全量替换文档
- update:执行的partial update操作
bulk api对json的语法,有严格的要求,每个json串不能换行,只能放一行,同时一个json串和一个json串之间,必须有一个换行
#语法错误
{
"error": {
"root_cause": [
{
"type": "json_e_o_f_exception",
"reason": "Unexpected end-of-input: expected close marker for Object (start marker at [Source: org.elasticsearch.transport.netty4.ByteBufStreamInput@5a5932cd; line: 1, column: 1])\n at [Source: org.elasticsearch.transport.netty4.ByteBufStreamInput@5a5932cd; line: 1, column: 3]"
}
],
"type": "json_e_o_f_exception",
"reason": "Unexpected end-of-input: expected close marker for Object (start marker at [Source: org.elasticsearch.transport.netty4.ByteBufStreamInput@5a5932cd; line: 1, column: 1])\n at [Source: org.elasticsearch.transport.netty4.ByteBufStreamInput@5a5932cd; line: 1, column: 3]"
},
"status": 500
}
执行结果
{
"took": 262,
"errors": false,
"items": [
{
"delete": {
"found": true,
"_index": "waws_index",
"_type": "waws_type",
"_id": "1",
"_version": 2,
"result": "deleted",
"_shards": {
"total": 2,
"successful": 1,
"failed": 0
},
"status": 200
}
},
{
"create": {
"_index": "waws_index",
"_type": "waws_type",
"_id": "3",
"_version": 1,
"result": "created",
"_shards": {
"total": 2,
"successful": 1,
"failed": 0
},
"created": true,
"status": 201
}
},
{
"index": {
"_index": "waws_index",
"_type": "waws_type",
"_id": "4",
"_version": 1,
"result": "created",
"_shards": {
"total": 2,
"successful": 1,
"failed": 0
},
"created": true,
"status": 201
}
},
{
"update": {
"_index": "waws_index",
"_type": "waws_type",
"_id": "3",
"_version": 2,
"result": "updated",
"_shards": {
"total": 2,
"successful": 1,
"failed": 0
},
"status": 200
}
}
]
}
bulk操作中,任意一个操作失败,是不会影响其他的操作的,但是在返回结果里,会告诉你异常日志
- 同一个index下的操作
POST /waws_index/_bulk
{ "delete": { "_type": "waws_type", "_id": "3" }}
{ "create": { "_type": "waws_type", "_id": "12" }}
{ "waws_name": "waws12" }
{ "index": { "_type": "waws_type" }}
{ "waws_name": "auto-generate id test" }
{ "index": { "_type": "waws_type", "_id": "2" }}
{ "waws_name": "replaced test2" }
{ "update": { "_type": "waws_type", "_id": "1", "_retry_on_conflict" : 3} }
{ "doc" : {"waws_name" : "waws_new_name"} }
POST /waws_index/waws_type/_bulk
{ "delete": { "_id": "3" }}
{ "create": { "_id": "12" }}
{ "waws_name": "waws12" }
{ "index": { }}
{ "waws_name": "auto-generate id test" }
{ "index": { "_id": "2" }}
{ "waws_name": "replaced test2" }
{ "update": { "_id": "1", "_retry_on_conflict" : 3} }
{ "doc" : {"waws_field" : "waws_new_name"} }
bulk size最佳大小
bulk request会加载到内存里,如果太大的话,性能反而会下降,因此需要反复尝试一个最佳的bulk size
。一般从1000-5000条数据开始,尝试逐渐增加。另外,如果看大小的话,最好是在5-15MB之间。