请尝试以下操作,可能并不理想,但会返回jobID和已用时间。首先我假设jobID
和moduleCode
是not_analyzed
:
{
"mappings": {
"jobs": {
"properties": {
"jobID":{
"type": "string",
"index": "not_analyzed"
},
"startTime":{
"type": "date"
},
"endTime":{
"type": "date"
},
"moduleCode":{
"type": "string",
"index": "not_analyzed"
}
}
}
}
}
用于scripted_metric
aggregation在ES 1.4.0可用于计算这两个值之间的差异。还没有研究如何添加过滤为“低于2880万”,但我希望可以有一些与该脚本进行限制这样的:
{
"query": {
"match_all": {}
},
"aggs": {
"jobIds": {
"terms": {
"field": "jobID"
},
"aggs": {
"executionTimes": {
"scripted_metric": {
"init_script": "_agg['time'] = 0L",
"map_script": "if (doc['moduleCode'].value == \"STARTING_MODULE\") { _agg['time']=-1*doc['startTime'].value } else { _agg['time']=doc['endTime'].value}",
"combine_script": "execution = 0; for (t in _agg.time) { execution += t };return execution",
"reduce_script": "execution = 0; for (a in _aggs) { execution += a }; return execution"
}
}
}
}
}
}
而且结果应该是这样的:
"aggregations": {
"jobIds": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": "ace4c888-1907-4021-a808-4a816e99aa1e",
"doc_count": 2,
"executionTimes": {
"value": 1
}
},
{
"key": "ace4c888-1907-4021-a808-4a816e99aa2e",
"doc_count": 2,
"executionTimes": {
"value": 1000201063
}
},
{
"key": "ace4c888-1907-4021-a808-4a816e99aa3e",
"doc_count": 2,
"executionTimes": {
"value": 10000
}
}
]
}
}