Amazon web services elasticsearch仅显示1个文档。使用logstash进行数据迁移
我正在尝试使用logstash和自定义模板将数据从S3(.csv文件的数据)移动到弹性搜索集群。 但当我在Kibana中使用以下查询进行检查时,它只显示docs.count=1,其余记录显示为docs.deleted:-Amazon web services elasticsearch仅显示1个文档。使用logstash进行数据迁移,amazon-web-services,elasticsearch,logstash,kibana,Amazon Web Services,elasticsearch,Logstash,Kibana,我正在尝试使用logstash和自定义模板将数据从S3(.csv文件的数据)移动到弹性搜索集群。 但当我在Kibana中使用以下查询进行检查时,它只显示docs.count=1,其余记录显示为docs.deleted:- GET /_cat/indices?v 我的第一个问题是:- 为什么只传输一条记录[最后一条]而其他记录则作为删除记录传输 现在,当我在Kibana中使用以下查询查询此索引时:- GET /my_file_index/_search { "query": { "m
GET /_cat/indices?v
我的第一个问题是:-
GET /my_file_index/_search
{
"query": {
"match_all": {}
}
}
我在“message:
字段中只得到一条数据以逗号分隔的记录,因此第二个问题是:-
columns => ["col1", "col2",...]
任何帮助都将不胜感激
编辑-1:下面是我的logstash.conf文件:-
input {
s3{
access_key_id => "xxx"
secret_access_key => "xxxx"
region => "eu-xxx-1"
bucket => "xxxx"
prefix => "abc/stocks_03-jul-2018.csv"
}
}
filter {
csv {
separator => ","
columns => ["AAA","BBB","CCC"]
}
}
output {
amazon_es {
index => "my_r_index"
document_type => "my_r_index"
hosts => "vpc-totemdev-xxxx.eu-xxx-1.es.amazonaws.com"
region => "eu-xxxx-1"
aws_access_key_id => 'xxxxx'
aws_secret_access_key => 'xxxxxx+xxxxx'
document_id => "%{id}"
template => "templates/template_2.json"
template_name => "my_r_index"
}
}
注:
logstash版本:6.3.1
elasticsearch版本:6.2
编辑:-2添加模板_2.json文件以及示例csv标题:-
1.映射文件:-
头:实际页眉很长,有很多栏,请继续考虑下面类似的其他列名。
SECURITY | SERVICEID | MEMBERID | VALUEDATE...
第一行:同样列值如下一些列有空白值,我提到了上面的真实模板文件(在上面的映射文件中),它有所有列值
KKK-LMN 2 1815 6/25/2018PPL-ORL 21815 6/25/2018
SLB-ORD 2 1815 6/25/2018
查询:
GET /my_r_index/_search
{
"query": {
"match_all": {}
}
}
输出:
{
"_index": "my_r_index",
"_type": "my_r_index",
"_id": "IjjIZWUBduulDsi0vYot",
"_score": 1,
"_source": {
"@version": "1",
"message": "XXX-XXX-XXX-USD,2,3190,2018-07-03,UNITED STATES,USD,300,60,Put,2042-12-19,,,,.009108041,q,,,,.269171754,q,,,,,.024127966,q,,,,68.414017367,q,,,,.298398645,q,,,,.502677959,q,,,,,0.040880692400344164,q,,,,,,,159.361792143,,,,.631296636,q,,,,.154877384,q,,42.93,N,Y,\n",
"@timestamp": "2018-08-23T07:56:06.515Z"
}
},
…上述其他类似记录
编辑-3:使用autodetect_列_name=>true后的示例输出:-
{
"took": 4,
"timed_out": false,
"_shards": {
"total": 10,
"successful": 10,
"skipped": 0,
"failed": 0
},
"hits": {
"total": 3,
"max_score": 1,
"hits": [
{
"_index": "indr",
"_type": "logs",
"_id": "hAF1aWUBS_wbCH7ZG4tW",
"_score": 1,
"_source": {
"2": "2",
"1815": "1815",
"message": """
PPL-ORD-XNYS-USD,2,1815,6/25/2018,UNITED STATES
""",
"SLB-ORD-XNYS-USD": "PPL-ORD-XNYS-USD",
"6/25/2018": "6/25/2018",
"@timestamp": "2018-08-24T01:03:26.436Z",
"UNITED STATES": "UNITED STATES",
"@version": "1"
}
},
{
"_index": "indr",
"_type": "logs",
"_id": "kP11aWUBctDorPcGHICS",
"_score": 1,
"_source": {
"2": "2",
"1815": "1815",
"message": """
SLBUSD,2,1815,4/22/2018,UNITEDSTATES
""",
"SLB-ORD-XNYS-USD": "SLBUSD",
"6/25/2018": "4/22/2018",
"@timestamp": "2018-08-24T01:03:26.436Z",
"UNITED STATES": "UNITEDSTATES",
"@version": "1"
}
},
{
"_index": "indr",
"_type": "logs",
"_id": "j_11aWUBctDorPcGHICS",
"_score": 1,
"_source": {
"2": "SERVICE",
"1815": "CLIENT",
"message": """
UNDERLYING,SERVICE,CLIENT,VALUATIONDATE,COUNTRY
""",
"SLB-ORD-XNYS-USD": "UNDERLYING",
"6/25/2018": "VALUATIONDATE",
"@timestamp": "2018-08-24T01:03:26.411Z",
"UNITED STATES": "COUNTRY",
"@version": "1"
}
}
]
}
}
我非常确定您的单个文档的id为
%{id}
。第一个问题来自这样一个事实:在CSV文件中,您没有提取一个名为id
的列,这就是您在document\u id=>“%{id}”
中使用的列,因此所有行都使用id%{id}
进行索引,并且每个索引都会删除前面的行。最后,您有一个文档,它的索引次数与CSV中的行数相同
关于第二个问题,您需要修复过滤器部分,如下所示:
filter {
csv {
separator => ","
autodetect_column_names => true
}
date {
match => [ "VALUATIONDATE", "M/dd/yyyy" ]
}
}
您还需要像这样修复索引模板(我只在估价日期
字段中添加了格式
设置:
{
"order": 0,
"template": "helloindex",
"settings": {
"index": {
"codec": "best_compression",
"refresh_interval": "60s",
"number_of_shards": "10",
"number_of_replicas": "1"
}
},
"mappings": {
"_default_": {
"_all": {
"enabled": false
},
"properties": {
"UNDERLYING": {
"type": "keyword"
},
"SERVICE": {
"type": "integer"
},
"CLIENT": {
"type": "integer"
},
"VALUATIONDATE": {
"type": "date",
"format": "MM/dd/yyyy"
},
"COUNTRY": {
"type": "keyword"
}
}
}
},
"aliases": {}
}
我很确定您的单个文档的id为
%{id}
,对吗?问题是,在CSV中,您没有提取名为id
的列,而这正是您在文档中使用的“%{id}”
因此,除最后一行外,所有行都被删除。您能显示CSV的真实标题和样本行吗?您是正确的,我必须删除从样本logstash.conf文件复制的文档“%{id}”,因为我在CSV标题中没有任何id列。@Val但是,我仍然在“message”中获得数据:“from not like in separate column field values。理想情况下,它应该在kibana输出中显示具有不同列名的值,对吗?我还为template_2.json文件中的所有列定义了映射。请随意共享尽可能多的信息,例如,您的模板、CSV文件的前2-3行,等等。实际上,它是一个CSV文件,所以我正在使用”作为分隔符。在我的问题中,我只使用|分隔标题列名。很抱歉造成混淆。请准确显示文件中出现的前两行,并使用我已更新答案的代码段的格式,管道在我这端使用您的示例数据。只是我获取的列值不是用于rec的标题ord 2 it从标题中pics,然后对于记录3 it从记录2中pics值作为标题,依此类推。由于logstash中的一些错误,这是由于同一文件上的并行工作线程,每个线程开始从不同的记录位置读取数据。解决方案是将--pipeline.workers设为1,以便尝试这样做,但不知何故它停止了数据传输为了索引.checking againNow,我将--pipeline.workers增加到了10,它传输了数据,但列标题仍然不匹配。我附上了一个示例输出,并附带了一个问题以供查看。
{
"order": 0,
"template": "helloindex",
"settings": {
"index": {
"codec": "best_compression",
"refresh_interval": "60s",
"number_of_shards": "10",
"number_of_replicas": "1"
}
},
"mappings": {
"_default_": {
"_all": {
"enabled": false
},
"properties": {
"UNDERLYING": {
"type": "keyword"
},
"SERVICE": {
"type": "integer"
},
"CLIENT": {
"type": "integer"
},
"VALUATIONDATE": {
"type": "date",
"format": "MM/dd/yyyy"
},
"COUNTRY": {
"type": "keyword"
}
}
}
},
"aliases": {}
}