我们之前看见了在Elasticsearch里的ingest node里,我们可以通过以下processor的处理帮我们处理我们的一些数据。它们的功能是非常具体而明确的。那么在Elasticsearch里,有没有一种更加灵活的方式可供我们来进行编程处理呢?如果有,它使用的语言是什么呢?
在Elasticsearc中,它使用了一个叫做Painless的语言。它是专门为Elasticsearch而建立的。Painless是一种简单,安全的脚本语言,专为与Elasticsearch一起使用而设计。 它是Elasticsearch的默认脚本语言,可以安全地用于inline和stored脚本。它具有像Groovy那样的语法。自Elasticsearch 6.0以后的版本不再支持Groovy,Javascript及Python语言。
脚本的语法为:
"script": {
"lang": "...",
"source" | "id": "...",
"params": { ... }
}
首先我们来创建一个简单的文档:
PUT twitter/_doc/1
{
"user" : "双榆树-张三",
"message" : "今儿天气不错啊,出去转转去",
"uid" : 2,
"age" : 20,
"city" : "北京",
"province" : "北京",
"country" : "中国",
"address" : "中国北京市海淀区",
"location" : {
"lat" : "39.970718",
"lon" : "116.325747"
}
}
在这个文档里,我们现在想把age修改为30,那么一种办法就是把所有的文档内容都读出来,让修改其中的age想为30,再重新用同样的方法写进去。首先这里需要有几个动作:先读出数据,然后修改,再次写入数据。显然这样比较麻烦。在这里我们可以直接使用Painless语言直接进行修改:
POST twitter/_update/1
{
"script": {
"source": "ctx._source.age = 30"
}
}
这里的source表明是我们的Painless代码。这里我们只写了很少的代码在DSL之中。这种代码称之为inline。在这里我们直接通过ctx._source.age来访问 _souce里的age。这样我们通过编程的办法直接对年龄进行了修改。运行的结果是:
{
"_index" : "twitter",
"_type" : "_doc",
"_id" : "1",
"_version" : 16,
"_seq_no" : 20,
"_primary_term" : 1,
"found" : true,
"_source" : {
"user" : "双榆树-张三",
"message" : "今儿天气不错啊,出去转转去",
"uid" : 2,
"age" : 30,
"city" : "北京",
"province" : "北京",
"country" : "中国",
"address" : "中国北京市海淀区",
"location" : {
"lat" : "39.970718",
"lon" : "116.325747"
}
}
}
显然这个age已经改变为30。上面的方法固然好,但是每次执行scripts都是需要重新进行编译的。编译好的script可以cache并供以后使用。上面的script如果是改变年龄的话,需要重新进行编译。一种更好的方法是改为这样的:
POST twitter/_update/1
{
"script": {
"source": "ctx._source.age = params.value",
"params": {
"value": 34
}
}
}
这样,我们的script的source是不用改变的,只需要编译一次。下次调用的时候,只需要修改params里的参数即可。
在Elasticsearch里:
"script": {
"source": "ctx._source.num_of_views += 2"
}
和
"script": {
"source": "ctx._source.num_of_views += 3"
}
被视为两个不同的脚本,需要分别进行编译,所以最好的办法是使用params来传入参数。
在这种情况下,scripts可以被存放于一个集群的状态中。它之后可以通过ID进行调用:
PUT _scripts/add_age
{
"script": {
"lang": "painless",
"source": "ctx._source.age += params.value"
}
}
在这里,我们定义了一个叫做add_age的script。它的作用就是帮我们把source里的age加上一个数值。我们可以在之后调用它:
POST twitter/_update/1
{
"script": {
"id": "add_age",
"params": {
"value": 2
}
}
}
通过上面的执行,我们可以看到,age将会被加上2。
Painless中用于访问字段值的语法取决于上下文。在Elasticsearch中,有许多不同的Plainless上下文。就像那个链接显示的那样,Plainless上下文包括:ingest processor, update, update by query, sort,filter等等。
Context | 访问字段 |
---|---|
Ingest node: 访问字段使用ctx | ctx.field_name |
Updates: 使用_source 字段 | ctx._source.field_name |
这里的updates包括_update,_reindex以及update_by_query。这里,我们对于context(上下文的理解)非常重要。它的意思是针对不同的API,在使用中ctx所包含的字段是不一样的。在下面的例子中,我们针对一些情况来做具体的分析。
首先我们创建一个叫做add_field_c的pipeline。关于如何创建一个pipleline,大家可以参考我之前写过的一个文章“如何在Elasticsearch中使用pipeline API来对事件进行处理”。
PUT _ingest/pipeline/add_field_c
{
"processors": [
{
"script": {
"lang": "painless",
"source": "ctx.field_c = (ctx.field_a + ctx.field_b) * params.value",
"params": {
"value": 2
}
}
}
]
}
这个pipepline的作用是创建一个新的field:field_c。它的结果是field_a及field_b的和,并乘以2。那么我们创建一个如下的文档:
PUT test_script/_doc/1?pipeline=add_field_c
{
"field_a": 10,
"field_b": 20
}
在这里,我们使用了pipleline add_field_c。执行后的结果是:
{
"took" : 147,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 1,
"relation" : "eq"
},
"max_score" : 1.0,
"hits" : [
{
"_index" : "test_script",
"_type" : "_doc",
"_id" : "1",
"_score" : 1.0,
"_source" : {
"field_c" : 60,
"field_a" : 10,
"field_b" : 20
}
}
]
}
}
显然,我们可以看到field_c被成功创建了。
在ingest过程中,可以使用脚本处理器来处理metadata,如_index和_type。 下面是一个Ingest Pipeline的示例,无论原始索引请求中提供了什么,它都会将索引和类型重命名为my_index:
PUT _ingest/pipeline/my_index
{
"description": "use index:my_index and type:_doc",
"processors": [
{
"script": {
"source": """
ctx._index = 'my_index';
ctx._type = '_doc';
"""
}
}
]
}
使用上面的pipeline,我们可以尝试index一个文档到any_index:
PUT any_index/_doc/1?pipeline=my_index
{
"message": "text"
}
显示的结果是:
{
"_index": "my_index",
"_type": "_doc",
"_id": "1",
"_version": 1,
"result": "created",
"_shards": {
"total": 2,
"successful": 1,
"failed": 0
},
"_seq_no": 89,
"_primary_term": 1,
}
也就是说真正的文档时存到my_index之中,而不是any_index。
PUT _ingest/pipeline/blogs_pipeline
{
"processors": [
{
"script": {
"source": """
if (ctx.category == "") {
ctx.category = "None"
}
"""
}
}
]
}
我们上面定义了一个pipeline,它可以帮我们检查如果 category字段是否为空,如果是,就修改为“None”。还是以之前的那个test_script索引为例:
PUT test_script/_doc/2?pipeline=blogs_pipeline
{
"field_a": 5,
"field_b": 10,
"category": ""
}
GET test_script/_doc/2
显示的结果是:
{
"_index" : "test_script",
"_type" : "_doc",
"_id" : "2",
"_version" : 2,
"_seq_no" : 6,
"_primary_term" : 1,
"found" : true,
"_source" : {
"field_a" : 5,
"field_b" : 10,
"category" : "None"
}
}
显然,它把category为“”的字段变为“None”了。
POST _reindex
{
"source": {
"index": "blogs"
},
"dest": {
"index": "blogs_fixed"
},
"script": {
"source": """
if (ctx._source.category == "") {
ctx._source.category = "None"
}
"""
}
}
上面的这个例子在reindex时,如果category为空时,写入“None”。我们可以从上面的两个例子中看出来,针对pipeline,我们可以直接对cxt.field进行操作,而针对update来说,我们可以对cxt._source下的字段进行操作。这也是之前提到的上下文的区别。
PUT test/_doc/1
{
"counter" : 1,
"tags" : ["red"]
}
您可以使用和update脚本将tag添加到tags列表(这只是一个列表,因此即使存在标记也会添加):
POST test/_update/1
{
"script" : {
"source": "ctx._source.tags.add(params.tag)",
"lang": "painless",
"params" : {
"tag" : "blue"
}
}
}
显示结果:
GET test/_doc/1
{
"_index" : "test",
"_type" : "_doc",
"_id" : "1",
"_version" : 4,
"_seq_no" : 3,
"_primary_term" : 11,
"found" : true,
"_source" : {
"counter" : 1,
"tags" : [
"red",
"blue"
]
}
}
显示“blue”,已经被成功加入到tags列表之中了。
您还可以从tags列表中删除tag。 删除tag的Painless函数采用要删除的元素的数组索引。 为避免可能的运行时错误,首先需要确保tag存在。 如果列表包含tag的重复项,则此脚本只删除一个匹配项。
POST test/_update/1
{
"script": {
"source": "if (ctx._source.tags.contains(params.tag)) { ctx._source.tags.remove(ctx._source.tags.indexOf(params.tag)) }",
"lang": "painless",
"params": {
"tag": "blue"
}
}
}
GET test/_doc/1
显示结果:
{
"_index" : "test",
"_type" : "_doc",
"_id" : "1",
"_version" : 5,
"_seq_no" : 4,
"_primary_term" : 11,
"found" : true,
"_source" : {
"counter" : 1,
"tags" : [
"red"
]
}
}
“blue”显然已经被删除了。
为了说明Painless的工作原理,让我们将一些曲棍球统计数据加载到Elasticsearch索引中:
PUT hockey/_bulk?refresh
{"index":{"_id":1}}
{"first":"johnny","last":"gaudreau","goals":[9,27,1],"assists":[17,46,0],"gp":[26,82,1],"born":"1993/08/13"}
{"index":{"_id":2}}
{"first":"sean","last":"monohan","goals":[7,54,26],"assists":[11,26,13],"gp":[26,82,82],"born":"1994/10/12"}
{"index":{"_id":3}}
{"first":"jiri","last":"hudler","goals":[5,34,36],"assists":[11,62,42],"gp":[24,80,79],"born":"1984/01/04"}
{"index":{"_id":4}}
{"first":"micheal","last":"frolik","goals":[4,6,15],"assists":[8,23,15],"gp":[26,82,82],"born":"1988/02/17"}
{"index":{"_id":5}}
{"first":"sam","last":"bennett","goals":[5,0,0],"assists":[8,1,0],"gp":[26,1,0],"born":"1996/06/20"}
{"index":{"_id":6}}
{"first":"dennis","last":"wideman","goals":[0,26,15],"assists":[11,30,24],"gp":[26,81,82],"born":"1983/03/20"}
{"index":{"_id":7}}
{"first":"david","last":"jones","goals":[7,19,5],"assists":[3,17,4],"gp":[26,45,34],"born":"1984/08/10"}
{"index":{"_id":8}}
{"first":"tj","last":"brodie","goals":[2,14,7],"assists":[8,42,30],"gp":[26,82,82],"born":"1990/06/07"}
{"index":{"_id":39}}
{"first":"mark","last":"giordano","goals":[6,30,15],"assists":[3,30,24],"gp":[26,60,63],"born":"1983/10/03"}
{"index":{"_id":10}}
{"first":"mikael","last":"backlund","goals":[3,15,13],"assists":[6,24,18],"gp":[26,82,82],"born":"1989/03/17"}
{"index":{"_id":11}}
{"first":"joe","last":"colborne","goals":[3,18,13],"assists":[6,20,24],"gp":[26,67,82],"born":"1990/01/30"}
文档里的值可以通过一个叫做doc的Map值来访问。例如,以下脚本计算玩家的总进球数。 此示例使用类型int和for循环。
GET hockey/_search
{
"query": {
"function_score": {
"script_score": {
"script": {
"lang": "painless",
"source": """
int total = 0;
for (int i = 0; i < doc['goals'].length; ++i) {
total += doc['goals'][i];
}
return total;
"""
}
}
}
}
}
这里我们通过script来计算每个文档的_score。通过script把每个运动员的goal都加起来,并形成最终的_score。这里我们通过doc['goals']这个Map类型来访问我们的字段值。显示的结果为:
{
"took" : 25,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 11,
"relation" : "eq"
},
"max_score" : 87.0,
"hits" : [
{
"_index" : "hockey",
"_type" : "_doc",
"_id" : "2",
"_score" : 87.0,
"_source" : {
"first" : "sean",
"last" : "monohan",
"goals" : [
7,
54,
26
],
"assists" : [
11,
26,
13
],
"gp" : [
26,
82,
82
],
"born" : "1994/10/12"
}
},
...
或者,您可以使用script_fields而不是function_score执行相同的操作:
GET hockey/_search
{
"query": {
"match_all": {}
},
"script_fields": {
"total_goals": {
"script": {
"lang": "painless",
"source": """
int total = 0;
for (int i = 0; i < doc['goals'].length; ++i) {
total += doc['goals'][i];
}
return total;
"""
}
}
}
}
显示的结果为:
{
"took" : 5,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 11,
"relation" : "eq"
},
"max_score" : 1.0,
"hits" : [
{
"_index" : "hockey",
"_type" : "_doc",
"_id" : "1",
"_score" : 1.0,
"fields" : {
"total_goals" : [
37
]
}
},
{
"_index" : "hockey",
"_type" : "_doc",
"_id" : "2",
"_score" : 1.0,
"fields" : {
"total_goals" : [
87
]
}
},
...
以下示例使用Painless脚本按其组合的名字和姓氏对玩家进行排序。 使用doc ['first']。value和doc ['last']。value访问名称。
GET hockey/_search
{
"query": {
"match_all": {}
},
"sort": {
"_script": {
"type": "string",
"order": "asc",
"script": {
"lang": "painless",
"source": "doc['first.keyword'].value + ' ' + doc['last.keyword'].value"
}
}
}
}
doc ['field'].value。如果文档中缺少该字段,则抛出异常。
要检查文档是否缺少值,可以调用doc ['field'] .size()== 0。
您还可以轻松更新字段。 您可以使用ctx._source.
首先,让我们通过提交以下请求来查看玩家的源数据:
GET hockey/_search
{
"stored_fields": [
"_id",
"_source"
],
"query": {
"term": {
"_id": 1
}
}
}
显示的结果为:
{
"took" : 0,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 1,
"relation" : "eq"
},
"max_score" : 1.0,
"hits" : [
{
"_index" : "hockey",
"_type" : "_doc",
"_id" : "1",
"_score" : 1.0,
"_source" : {
"first" : "johnny",
"last" : "gaudreau",
"goals" : [
9,
27,
1
],
"assists" : [
17,
46,
0
],
"gp" : [
26,
82,
1
],
"born" : "1993/08/13"
}
}
]
}
}
要将玩家1的姓氏更改为hockey,只需将ctx._source.last设置为新值:
POST hockey/_update/1
{
"script": {
"lang": "painless",
"source": "ctx._source.last = params.last",
"params": {
"last": "hockey"
}
}
}
您还可以向文档添加字段。 例如,此脚本添加一个包含玩家nickname为hockey的新字段。
POST hockey/_update/1
{
"script": {
"lang": "painless",
"source": """
ctx._source.last = params.last;
ctx._source.nick = params.nick
""",
"params": {
"last": "gaudreau",
"nick": "hockey"
}
}
}
显示的结果为:
GET hockey/_doc/1
{
"_index" : "hockey",
"_type" : "_doc",
"_id" : "1",
"_version" : 2,
"_seq_no" : 11,
"_primary_term" : 1,
"found" : true,
"_source" : {
"first" : "johnny",
"last" : "gaudreau",
"goals" : [
9,
27,
1
],
"assists" : [
17,
46,
0
],
"gp" : [
26,
82,
1
],
"born" : "1993/08/13",
"nick" : "hockey"
}
}
有一个叫做 “nick”的新字段被加入了。
我们甚至可以对日期类型来进行操作从而得到年月等信息:
GET hockey/_search
{
"script_fields": {
"birth_year": {
"script": {
"source": "doc.born.value.year"
}
}
}
}
显示结果:
{
"took" : 0,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 11,
"relation" : "eq"
},
"max_score" : 1.0,
"hits" : [
{
"_index" : "hockey",
"_type" : "_doc",
"_id" : "2",
"_score" : 1.0,
"fields" : {
"birth_year" : [
1994
]
}
},
...
Elasticsearch第一次看到一个新脚本,它会编译它并将编译后的版本存储在缓存中。无论是inline或是stored脚本都存储在缓存中。新脚本可以驱逐缓存的脚本。默认的情况下是可以存储100个脚本。我们可以通过设置script.cache.max_size来改变其大小,或者通过script.cache.expire来设置过期的时间。这些设置需要在config/elasticsearch.yml里设置。
不能调试的脚本是非常难的。有一个好的调试手段无疑对我们的脚本编程是非常有用的。
Painless没有REPL,虽然有一天它很好,但它不会告诉你关于调试Elasticsearch中嵌入的Painless脚本的全部故事,因为脚本可以访问的数据或“上下文” 是如此重要。 目前,调试嵌入式脚本的最佳方法是在选择位置抛出异常。 虽然您可以抛出自己的异常(throw new exception('whatever'),但Painless的沙箱会阻止您访问有用的信息,如对象的类型。 所以Painless有一个实用工具方法Debug.explain,它会为你抛出异常。 例如,您可以使用_explain来探索script query可用的上下文。
PUT /hockey/_doc/1?refresh
{"first":"johnny","last":"gaudreau","goals":[9,27,1],"assists":[17,46,0],"gp":[26,82,1]}
POST /hockey/_explain/1
{
"query": {
"script": {
"script": "Debug.explain(doc.goals)"
}
}
}
这表明doc.goals类是org.elasticsearch.index.fielddata.ScriptDocValues.Longs通过响应:
{
"error": {
"root_cause": [
{
"type": "script_exception",
"reason": "runtime error",
"painless_class": "org.elasticsearch.index.fielddata.ScriptDocValues.Longs",
"to_string": "[1, 9, 27]",
"java_class": "org.elasticsearch.index.fielddata.ScriptDocValues$Longs",
"script_stack": [
"Debug.explain(doc.goals)",
" ^---- HERE"
],
"script": "Debug.explain(doc.goals)",
"lang": "painless"
}
],
"type": "script_exception",
"reason": "runtime error",
"painless_class": "org.elasticsearch.index.fielddata.ScriptDocValues.Longs",
"to_string": "[1, 9, 27]",
"java_class": "org.elasticsearch.index.fielddata.ScriptDocValues$Longs",
"script_stack": [
"Debug.explain(doc.goals)",
" ^---- HERE"
],
"script": "Debug.explain(doc.goals)",
"lang": "painless",
"caused_by": {
"type": "painless_explain_error",
"reason": null
}
},
"status": 400
}
您可以使用相同的技巧来查看_source是_update API中的LinkedHashMap:
POST /hockey/_update/1
{
"script": "Debug.explain(ctx._source)"
}
显示的结果是:
{
"error": {
"root_cause": [
{
"type": "remote_transport_exception",
"reason": "[localhost][127.0.0.1:9300][indices:data/write/update[s]]"
}
],
"type": "illegal_argument_exception",
"reason": "failed to execute script",
"caused_by": {
"type": "script_exception",
"reason": "runtime error",
"painless_class": "java.util.LinkedHashMap",
"to_string": "{first=johnny, last=gaudreau, goals=[9, 27, 1], assists=[17, 46, 0], gp=[26, 82, 1], born=1993/08/13, nick=hockey}",
"java_class": "java.util.LinkedHashMap",
"script_stack": [
"Debug.explain(ctx._source)",
" ^---- HERE"
],
"script": "Debug.explain(ctx._source)",
"lang": "painless",
"caused_by": {
"type": "painless_explain_error",
"reason": null
}
}
},
"status": 400
}
参考:
【1】https://www.elastic.co/guide/en/elasticsearch/painless/current/painless-walkthrough.html
【2】https://www.elastic.co/guide/en/elasticsearch/painless/current/painless-debugging.html