find on binData

[Mike Atkins]

I've been experimenting with using a custom _id field for a gridfs fs.files collection so I can use files_id in fs.chunks as a sharding key. I am using a byte array consisting of two bytes of random data followed by 12 bytes from an ObjectId. Seems to do what I want in terms of sharding, although I am aware that having the _id field randomized can cause a performance hit related to index caching. In any case, I was surprised to see the following (am I not using the right syntax?): 

mongos> db.fs.files.find({}, {_id: 1})

{ "_id" : BinData(0,"lZZOnIbU+ZJkA/7YpAw=") }

{ "_id" : BinData(0,"gpJOnIbX+ZJkAwTZpAw=") }

{ "_id" : BinData(0,"E2NOnIbf+ZJkAwrZpAw=") }

mongos> db.fs.files.find({_id: BinData(0,"lZZOnIbU+ZJkA/7YpAw=") }, {_id: 1})

{ "_id" : BinData(0,"lZZOnIbU+ZJkA/7YpAw=") }

{ "_id" : BinData(0,"gpJOnIbX+ZJkAwTZpAw=") }

{ "_id" : BinData(0,"E2NOnIbf+ZJkAwrZpAw=") }

mongos> db.fs.files.find({_id: BinData(0,"lZZOnIbU+ZJkA/7YpAw=") }, {_id: 1}).count()

0

At the very least the second two results seem inconsistent, although I think both are wrong (I was expecting one document to be returned).

what am I missing?

I'm using version 2.0.0.

Thanks,

-Mike

[Mike Atkins]

Looking at this some more, I think this is a bug. A the very least an inconsistency between the way HexData and BinData work. 

> db.bindata.insert({ foo: BinData(0,"gpJOnIbX+ZJkAwTZpAw=")})

> db.bindata.find()

{ "_id" : ObjectId("4e9d87aa5825b60b637815a6"), "foo" : null }

so BinData(0,"gpJOnIbX+ZJkAwTZpAw=")produces null..

> db.bindata.insert({ foo: HexData(0,"5555")})

> db.bindata.find()

{ "_id" : ObjectId("4e9d87aa5825b60b637815a6"), "foo" : null }

{ "_id" : ObjectId("4e9d87d15825b60b637815a7"), "foo" : BinData(0,"VVU=") }

whereas HexData(0, "5555") produces a non-null value, as expected.

If I use the new keyword with BinData, it does what I expect:

> db.bindata.insert({ foo: new BinData(0,"gpJOnIbX+ZJkAwTZpAw=")})

> db.bindata.find()

{ "_id" : ObjectId("4e9d87aa5825b60b637815a6"), "foo" : null }

{ "_id" : ObjectId("4e9d87d15825b60b637815a7"), "foo" : BinData(0,"VVU=") }

{ "_id" : ObjectId("4e9d87e15825b60b637815a8"), "foo" : BinData(0,"gpJOnIbX+ZJkAwTZpAw=") }

and..

> db.bindata.find({foo: new BinData(0,"gpJOnIbX+ZJkAwTZpAw=")})

{ "_id" : ObjectId("4e9d87e15825b60b637815a8"), "foo" : BinData(0,"gpJOnIbX+ZJkAwTZpAw=") }

so, basically "new" needed with BinData, but not HexData.

[Mike Atkins]

filed as https://jira.mongodb.org/browse/SERVER-4093

[Mathias Stearn]

Thanks for reporting this. Unfortunately JS semantics can be weird when it comes to constructing objects not implemented in JS, including built-in types like Array. We'll see what we can do to normalize the syntaxes.

[Grégoire Seux]

Hello,

this issue while searching with field of type bindata is quite embarassing.

For instance, I search a group of elements with the shard key (_id which is of BinData type).

my query is using $in : [.... array of binDatas]

some of the elements seems to be found and some (always the same number) are not.

How can I do a proper query to make sure to avoid this issue ? (I am using the c# driver 1.3 and mongo 2.0.0)

-- 

Greg

[Mathias Stearn]

I think you should be able to query using c#'s binary type (BsonBinaryData). If that isn't working for you make sure you are using the same subtype.

[Robert Stam]

Gregoire, not sure what you are asking...

If you could provide a working JSON query in the mongo shell and the same query in your C# program I can verify for you whether they are equivalent.

On Mon, Nov 21, 2011 at 7:12 PM, Mathias Stearn <mat...@10gen.com> wrote:

I think you should be able to query using c#'s binary type (BsonBinaryData). If that isn't working for you make sure you are using the same subtype.

--
You received this message because you are subscribed to the Google Groups "mongodb-user" group.

To view this discussion on the web visit https://groups.google.com/d/msg/mongodb-user/-/QKDDXrp4b7IJ.

To post to this group, send email to mongod...@googlegroups.com.
To unsubscribe from this group, send email to mongodb-user...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en.