angelmunoz/mondocks

An alternative way to interact with MongoDB databases from F# that allows you to use mongo-idiomatic constructs

mondocks's Issues

geoSearch Command

https://docs.mongodb.com/manual/reference/command/geoSearch/#dbcmd.geoSearch

Serialization errors with `FindResult`, `DeleteResult` on dotnet

Running the examples in samples/Commands fails because the MongoDB driver expects certain fields that don't exist in a FindResult<'T> or a UpdateResult to actually exist.

On my end, I can confirm that simply doing

[<BsonIgnoreExtraElements>]
type FindResult<'T> = { cursor: Cursor<'T>; ok: float }

resolves this (because Mongo also returns the operationTime field).

However, it's not just this one simple fix; the update command still fails because of the same issue. My conclusion is that we'll have to add this attribute to all of the types in Mondocks.Net's Types.fs; it follows that it might not make sense to share the types between Mondocks.Net and Mondocks.Fable?

Finally, a small feature request/design question: what determines which fields make it into the result structs? It would be nice, for instance, to also be able to also inspect the upserted items of an UpdateResult. See output reference here.

Create Indexes

https://docs.mongodb.com/manual/reference/command/createIndexes/#dbcmd.createIndexes

Drop Indexes

https://docs.mongodb.com/manual/reference/command/dropIndexes/#dbcmd.dropIndexes

Make it available for Node

Most of the builders don't use the Mongo.BsonId package it's basically used at the end to generate the final JSON, and it is basically used just to add some serializers but if we switch to Thoth.Json perhaps we should be able to drop that dependency and allow the library to be used from .NET and Node

Query for a subset of the document

It should be possible to define a document only partially and the query should still work.
At the moment I have to define every JSON field or else I get a FormatException.

Example:

Document

{
  "_id": "foo",
  "Active": true,
  "LicenseInfo": {
    ...
  },
  ...
}

Query

type Poi =
  { _id: string
    Active: bool }

let q = find "pois" { ... }
db.RunCommandAsync<FindResult<Poi>>(JsonCommand q)

System.FormatException: An error occurred while deserializing the cursor property of class Mondocks.Types.FindResult`1[[FSI_0063+Poi, FSI-ASSEMBLY, Version=0.0.0.0, Culture=neutral, PublicKeyToken=null]]: An error occurred while deserializing the firstBatch property of class Mondocks.Types.Cursor`1[[FSI_0063+Poi, FSI-ASSEMBLY, Version=0.0.0.0, Culture=neutral, PublicKeyToken=null]]: Element 'LicenseInfo' does not match any field or property of class FSI_0063+Poi. ---> System.FormatException: An error occurred while deserializing the firstBatch property of class Mondocks.Types.Cursor`1[[FSI_0063+Poi, FSI-ASSEMBLY, Version=0.0.0.0, Culture=neutral, PublicKeyToken=null]]: Element 'LicenseInfo' does not match any field or property of class FSI_0063+Poi. ---> System.FormatException: Element 'LicenseInfo' does not match any field or property of class FSI_0063+Poi.
   at MongoDB.Bson.Serialization.BsonClassMapSerializer`1.DeserializeClass(BsonDeserializationContext context)

Especially in a production environment, where the document schema can grow independently from your application code this ends up in runtime errors that can easily avoided by allowing to map to a subset of the document.

explore the posibility to add BSON decoders

taken from #1

Support BsonNull.Value to query for null or unset fields

mapReduce Command

https://docs.mongodb.com/manual/reference/command/mapReduce/#dbcmd.mapReduce

Add Docs

Add proper documentation explaining the use cases of this library as well it's computation expressions

Filter with int default (0) value.

With options.DefaultIgnoreCondition <- JsonIgnoreCondition.WhenWritingDefault settings expression like this:

find "users" {
    filter {| age = {| ``$gt`` = 0 |} |}
    limit 5
}

is producing this json:

{"find":"users","filter":{"age":{}},"limit":{"$numberInt":"5"}}

and we get no documents.

But with options.DefaultIgnoreCondition <- JsonIgnoreCondition.WhenWritingNull:

{"find":"users","filter":{"age":{"$gt":{"$numberInt":"0"}}},"limit":{"$numberInt":"5"}}

and filter works as expected.

I suppose you should replace the settings in default serializer options.

Affected version: Mondocs.Net (0.4.8)

Investigate helper functions

taken from #1
add method helpers to create things like filters, fields, matching criteria

examples of this can be

find "users" {
  filter (fun m -> m.prop = value)
}

or

find "users" {
  filter "prop" gt 10
}

the main blocker right now for this is that in any case, we must emit 100% compatible JSON to https://docs.mongodb.com/manual/reference/command/
so there needs to be some investigation around this if this is something people want to add into the mix

Add Tests

Since this library produces JSON testing should not be complicated at all we just need to check against an existing JSON for each command

Explore default BSON serialization

There's the possibility that we can just serialize known BSON values in the emitted JSON, as the MongoDB Extended Spec shows, we're kind of doing half work for ObjectIds so instead of just emitting the "objectid" we could be emitting { $oid: "objectid" } in the same way we could be doing the same for int64, dates, etc

Ideally I'd like to use the canonical form when possible, but if there's not an option using the relaxed form should be enough

• Date
• Decimal128
• Double
• Int32
• Int64
• ObjectId
• Regular Expression
• Binary
• MinKey
• MaxKey
• Array
• Timestamp

Feedback from twitter

I took a few notes from Twitter earlier today

• look a way to create indexes
• remove prefixes from custom operations
• pass the collection name as the parameter of the CE's
• return typed results from the command execution
• add BSON decoders
• add method helpers to create things like filters, fields, matching criteria

Distinct command

https://docs.mongodb.com/manual/reference/command/distinct/#dbcmd.distinct

Aggregate command

https://docs.mongodb.com/manual/reference/command/aggregate/#dbcmd.aggregate

Count command

https://docs.mongodb.com/manual/reference/command/count/#dbcmd.count

Example of using array filters

Hello!

Can someone show an example of using array filters in queries?
Looks like currently they are represented as just "object" type, so its not clear on how to actually utilize them.

Thank you!