Mongodb - multi-criteria search

Question

MongoDB

We need advice from those who worked with MongoDB:
Let's say we have data like this:

{
  a: {
    id:1,
    id2:3
  },	
  b: [ "str1", "str2", "str3" ],
  c: [ 
    {f:"UA", g:12, s:"some string"},
    {f:"RU", g:1, s:"some string2"}
   ],
  d: [ 
    {f:"UA", g:2, s:"some string3"},
    {f:"UA", g:2, s:"some string4"}
   ],
  dtF: ISODate("2012-04-25T21:00:00Z"),
  dtT: ISODate("2012-04-28T21:00:00Z"),
  cc: true,
  m: 0,
  v: 32,
  ...
}

There are about 3M records in the collection.

It is necessary to search this collection in different fields, i.e. in one request, all fields can participate in another, only a few.

Request examples:

{ "c.f" : "UA" , "d.f" : "UA" , "c.s":{"$regex":"some", "$options" : ""}} 

{ "c.f" : "UA" , "d.f" : "UA" , "c.g" : { "$in" : [ 17]}, "b":{"$nin":["str1"]} 

{ "m" : { "$gte" : 2 , "$lte" : 7} , "v" : { "$lte" : 50} , "c.f" : "UA" , "d.f" : "UA" , "c.g" : { "$in" : [ 27]}}

{"m" : { "$lte":20} , "v" : { "$lte" : 8},"c.f":"UA","c.g":{"$in":[1,12,16]}, "c.s":{"$regex":"some", "$options" : ""},"d.f":"RU","d.g":{"$in":[25,16]}, "d.s":{"$regex":"some2", "$options" : ""},cc:false, "$or" : [ { "dtF" : { "$gte" : { "$date" : "2012-05-06T21:00:00.000Z"}} , "dtT" :  null } , { "dtF" : { "$gte" : { "$date" : "2012-05-06T21:00:00.000Z"}} , "dtT" : { "$lte" : { "$date" : "2012-06-06T21:00:00.000Z"}}}], "b":{"$in":["str1"]}}

Problems that I see
: There can be many different query options
- You cannot create a composite index that includes several "array fields"
- The more indexes, the lower the write speed.

What do you advise? Is MongoDB up to the task?

Answer 1

I advise you to take a sample of the fields of the most popular requests and, based on these fields, build your own index for each request. Solve the performance problem by 80%

Answer 2

MongoDB is, of course, cool, web scale and all, but you seem to be solving the problem of parametric search. There are specially designed tools for just this task. Lucy, for example.