Question

我有一组过滤数据，如下格式

filters = [  
{key : colour, values: [red,blue,green,yellow,........]},  
{key:website, values:[myntra, jabong, voonik,...........]},  
{key:shape,values:[fit, maxi,bodycon, skater...................]}  
,........................  
,........................]

我的弹性搜索数据库结构是

{  
"_index": "products_data",  
"_type": "dresses",  
"_id": "1",  
"_score": 0,  
"_source": {  
"product_filter":{  
    "dress_shape": "sheath",  
    "pattern_type": "solid",  
    "discount_price": 1347,  
    "knit_or_woven": "knit",  
    "year": "2015",  
    "age_group": "adults-women",  
    "broad_category": "dress",  
    "fabric": "polyester",    
    "lining": "has a lining",  
    "surface_styling_or_features": "other",  
    "usage": "casual",  
    "sleeves_type": "sleeveless",  
    "brand": "deal jeans",  
    "website": "myntra",  
    "season": "fall",  
    "price": 2695,  
    "discount_percent": 50,  
    "product_line": "dresses",  
    "neck": "round neck",  
    "sleeve": "sleeveless",  
    "gender": "women",  
    "colour": "black",  
    "occasion": "casual",  
    "dress_length": "mini",  
    "display_name": "deal jeans black sheath dress",  
    "hemline": "curved",  
    "fabric_type": "lace or crochet"  
    }  
  }  
}

我需要找到每个过滤器计数，例如image。目前我正在采取每个过滤器并生成弹性搜索查询，如下面的格式，并将此查询发送到弹性搜索计数api进行计数。

{ "query": { "bool": { "filter":{ "term": { "product_filter.brand": "109f" } } } } }

output :  109f brand --> 2132

我在每个过滤器列表中有更多数据。计算500个过滤器大约需要6秒钟我尝试了多搜索api（msearch api），但它也花了很多时间。我的数据大小是19787033681字节，它有5个分片。任何人都可以通过节点js代码帮助我...

Answer 1

我的建议是将分片数量从5减少到1.除非您有多台机器处理相同的数据或希望使用群集，否则您应该只使用一个分片。

{
   "settings":{
       "number_of_shards":1,
       "number_of_replicas":0
   }
}

使用多个服务器/机器进行插入/搜索/删除操作和冗余时，使用多个分片和副本非常有效。

如何在节点js中以更少的时间获得以下弹性搜索数据库结构的结果？

1 个答案: