Logastash CSV过滤器|转换

从csv转换某些值时遇到问题。 如果我只声明一个，例如'id'=>'integer'，它可以正常工作。 如果我在下面的配置中添加另一个类似的转换，它将崩溃

    input {
        tcp {
            port => 5000
        }
        file {
              path => '/tmp/logstash/databases/1.csv'
              type => 'test'
              start_position => 'beginning'
          }
    }

    filter {
        csv {
            columns => ['Id','ElapsedRealTimeNanos','Accuracy','Altitude','Bearing','CreatedOn','Latitude','Longitude','Provider','Speed','Time']
            separator => ','
            convert => {'Accuracy' => 'float','id'=>'integer'}
        }
    }

    output {
        elasticsearch {
                hosts => 'elasticsearch:9200'
        }
        stdout{

        }
    }

错误：

{:timestamp=>"2016-04-27T18:49:51.881000+0000", 
:message=>"fetched an invalid config", 
:config=>"input {\n\ttcp {\n\t\tport => 5000\n\t}\n\tfile {\n 
path => '/tmp/logstash/databases/1.csv'\n          
type => 'test'\n          start_position => 'beginning'\n 
}\n}\n\nfilter {\n    csv {\n        
columns => ['Id','ElapsedRealTimeNanos','Accuracy','Altitude','Bearing','CreatedOn','Latitude','Longitude','Provider','Speed','Time']\n        
separator => ','\n\t\t\t\tconvert => {'Accuracy' => 'float','id'=>'integer'}\n    }\n}\n\noutput {\n\telasticsearch {\n\t\t\thosts => 'elasticsearch:9200'\n\t}\n\tstdout{\n\n\t}\n}\n\n", 
:reason=>"Expected one of #, {, } at line 16, column 38 (byte 391) after filter {\n    csv {\n       
columns => ['Id','ElapsedRealTimeNanos','Accuracy','Altitude','Bearing','CreatedOn','Latitude','Longitude','Provider','Speed','Time']\n        
separator => ','\n\t\t\t\tconvert => {'Accuracy' => 'float'", :level=>:error}

您可以将其用作：

...
convert => {
    'Accuracy' => 'float'
    'Id' => 'integer'
}
...

转换之间没有逗号

您还可以使用--configtest选项检查配置文件是否存在错误：

λ bin/logstash -f example.conf --configtest

---

谢谢阿尔伯特-不再是错误了。但不幸的是，字符串未被转换：/精度保存为30.0以便更清楚：在kibana中，它将“字符串”显示为accuracyDelete索引和logstash创建的索引模板的类型。从csv中删除标题。它会创建数字格式异常。重新编制数据索引。这对我很管用。另外，

id

必须是

id

。更新：使用mutate转换它有效！