Log file Viewer
#mongodb logs for Wednesday the 26th of March, 2014
[00:03:48] <ShortWave> I'm using a 2d query (with legacy coordinates, don't ask), and while that part works fine...what I need to do is also return the distance from a given record to a provided point.
[00:17:43] <Ramjet> HOW do I build the client libraries and includes... The scons --full used to do it but it has been DEPRECATED and I do not see any new targets in the SConstruct file that replaces --full ?????
[00:31:04] <Ramjet> HELP!!!!!! HOW do I build the client libraries and includes... The scons --full used to do it but it has been DEPRECATED and I do not see any new targets in the SConstruct file that replaces --full ?????
[08:56:51] <Guest92111> hi, I'm building a restful API on a mongo database, and am having troubles figuring out how to properly implement pagination over a collection, ordered by an updatedAt timestamp, anyone have some experience with this problem?
[08:59:38] <Guest92111> when i paginate by 50 results for example, when retrieving the second page, some documents in the first page might have been already updated, which affect the entire result set, I thougt about solving this by specifying a "since" timestamp which is set to the updatedAt of the last result in my current page, but a problem arises when different documents might have the same updatedAt timestamp
[09:01:57] <gerryvdm> yeah but say i retrieve 50 documents with the same timestamp, and there are 50 more with that timestamp in the next page, if I ask for all documents "since" the last updated at I keep looping over the first 50
[09:02:33] <gerryvdm> I actually was thinking about using an ObjectId as the column to sort on, would that be a solution?
[09:03:12] <gerryvdm> Nodex: in that case I no longer have a "previous" page but I can still retrieve the next set
[09:04:28] <gerryvdm> well i "was" at the updatedat timestamp of the last document in my previous page
[09:06:23] <gerryvdm> if i use ObjectId not only for the _id but also for a sort column, wouldnt that work? I'd create a new ObjectId on every update
[09:09:05] <gerryvdm> if i am in the middle of pagination, and i'm at everything updated since midnight last night, if something of a previous page changes right now, i would still get all changes and also my new change in the last page?
[09:09:40] <Nodex> [08:59:24] <Guest92111> when i paginate by 50 results for example, when retriev.......
[09:11:06] <gerryvdm> no say in my first page (by 2 results) i get two documents "updatedAt: 11:00, updatedAt: 12:00"
[09:11:31] <jinmatt> is this some kind of logging data that needs to be paginated, as per your description I have a feeling that its never possible to see the end of this list as your data entries keeps changing so fequent
[09:11:39] <gerryvdm> for my next query i need to ask the next 2 documents where updatedAt is greater than 12:00
[10:54:28] <jinmatt> mms agent is having problems to connect to ec2 instance after adding firewall security rules, I’m getting the follow error from the Agent Log “Problem collecting blocking data: could not connect to ip-xx-xx-xx-xx:27017: timed out”
[10:55:43] <jinmatt> otherwise even through I add the agent host machines’s IP address to access rule list on EC2, its not working
[11:38:40] <jinmatt> rules, I’m getting the follow error from the Agent Log “Problem collecting blocking data: could not connect to ip-xx-xx-xx-xx:27017: timed out”
[13:05:30] <Industrial> I'm trying to figure out what the best way to save my data is in mongodb. The data is time based so I'm using millisecond timestamps for the _id index of the collections, and per timestamp that happened there may be several keys set. What I'd like to do now is to split that into one collection per key kind instead.
[13:06:27] <Industrial> Because not every document may contain every key, when I make queries for certain key kinds, I'm also querying all data that does not contain those keys. I could eliminate that constraint by just querying a database for the specific data I'm looking for.
[13:07:24] <Industrial> With the Node.js driver, does calling `var collection = db.collection('test_insert');` do anything on the database end? no right? so I could potentially call this thousands of times per second?
[13:08:20] <Industrial> because I'm receiving a LOT of messages, not knowing which collection they should land in.
[13:14:28] <Nodex> Industrial : your "var" might have a local or global scope, depending on where you set it
[13:15:31] <Industrial> Nodex: yeah, but there's notghing going forth and back between mongo and node, trying to 'fetch' a collection right? It just modifies calls made to that object to work on a certain collection rather then another?
[13:16:49] <dexteruk> im new to the world of mongodb, i have a quick question, im looking at the best way to structure my documents, i have a set of dynamic elements but i have read about Embedded Documents that the document can grow which can impact write performance, is this really an issue, and what things should i consider
[13:17:21] <Industrial> dexteruk: try and get soemthing rolling on the system first, and then worry about performance
[13:21:29] <Nodex> dexteruk : you should think about your structure first. This is the exact opposite of other systems
[13:21:55] <Nodex> you should tailor your app to your data and the fastest way to read/write it else you will NEVER get good performance
[13:22:25] <Nodex> Industrial : the driver will communicate when there is something to send to the database
[13:23:16] <dexteruk> yes i have seen this, in some examples, where you put the content and embed the link
[14:38:38] <ShortWave> Is there a way to get a distance calculation from a 2d query, without doing a calculation myself?
[14:47:55] <yug> Does anyone know if this is possible to do aggregation on aggregation with aggregation module? i.e first to aggregate by hours and the result of this I want to aggregate by day
[14:53:28] <rybnik> yug would something like this do the work {$group: {_id: { h: {$hour: '$tick'}, d: {$dayOfMonth: '$tick'} }, n: {$sum:1}}} ?
[14:55:38] <yug> rybnik no because this will group by both together. I need to group first by hour and the result to group by day
[14:56:23] <rybnik> yug, sorry, could you illustrate with an example (of the expected result) - use a pastie please ?
[14:56:43] <yug> it's like putting it buckets all the data in hour resolution and then bucket by days
[15:01:52] <abique_> Hi, I'm using the C++ driver, and I wonder how to do a query "find by _id" is it conn.query("mydb", QUERY("_id" << "45645645645689765"));
[16:04:46] <palominoz> hello there. given my multilanguage documents look like { id: 'some-mongo-id', locales: { en: {type: '}, it: { name: "pane"}}}
[16:08:06] <palominoz> hello there. given my multilanguage documents look like { id: 'some-mongo-id', locales: { en: {type: 'bread'}, it: { name: "pane"}}}, i would like to support fulltext search based on locale, like db.myCollection.runCommand("text", {search: "pane", language: "italian" }). I did run ensureIndex for each language indexable in the locales subdocument. mongo then stops me with error: too many text index for: production.manuf
[16:08:06] <palominoz> acturers. how would you solve this problem? misspressed enter button before :)
[16:09:01] <palominoz> is it possible to have a single index and then discriminate language on the results?
[16:34:47] <LinkRage> When I use $group in Mongo aggregation the output is array named result. How do I print *only* the inner fields of that array in Mongo CLI?
[17:23:05] <paper_ziggurat> is it possible to iterate over a collection and then return a value where a certain json element is equal to a predetermine value?
[17:27:24] <saalaa__> hello, anyone has any experience accessing 'local.system.namespaces' on MongoLab?
[17:28:48] <saalaa__> (specifically I'm trying to install an Elastic Search MongoDB River which needs access to it but doesn't seem to be able to access it:)
[17:30:39] <paper_ziggurat> NaN, I have a key:value pair, I want to iterate over the collection and return the value where one of the value's json elements is equal to something
[17:48:20] <saalaa__> for those who might skim through the backlog, it turns out MongoDB's *.system.namespaces, which is needed for an Elastic Search River, is not available MongoLab's plan bellow "dedicated" offerings starting at 200 USD
[18:08:37] <paper_ziggurat> the query is supposed to find a document 'where' a json element in the document is a certain value
[19:01:28] <paper_ziggurat> i thought find where select meant that it selects a document where a value in the document is equal to some other value
[19:08:11] <paper_ziggurat> after running the query though, the value that's returned is a big JSON document with details about the collection
[19:33:29] <Guest43094> I’m using 2d indexes in a multiplayer game world to find players near my current position on the map that are online and of a certain type. I’d like to use a compound index for this, but this is obviously only possible with 2dsphere. What I’m not clear about is if the 2dspehere index will return different results than the 2d index for the same location/data.
[20:16:27] <NaN> when I do >new Date() it doesn't show my current GMT (I know it's ISO) but is not my current time, why?
[20:33:58] <slon> I have an unique index on multiple fields within a collection. When I insert a document containing duplicate data for multiple unique fields I only get back one error. Is there a way to get all errors back?
[20:35:28] <slon> I'm only getting back the first duplicate error, but I want to know all fields that are duplicates for the specific entry.
[20:36:18] <slon> like if name and email have unique separate unique indices and I'm adding a name and an email that both already exist, I want to get back errors about both fields. Currently, I'm only getting back errors for one field.
[21:16:03] <proteneer> question about the mongo license, if I build mongo from source, then am I allowed to use SSL?
[21:21:00] <slon> hey! I have a new document and I want to know which fields will be duplicates when inserted. Is there a way to get all duplicate field errors when inserting?
[21:22:49] <proteneer> it's stupid to not include SSL when its as simple as adding in --ssl as the build option
[21:26:06] <bob_123> hello, I have a question that seems silly but I can't get this to work for me: in a find query, how do you get mongo to match a document where a field value is 0?
[21:27:35] <ranman> slon: have you considered findAndModify()? you could probably design the functionality you want around that
[21:28:53] <bob_123> if I was doing update, it would be the same, right: collection.update( { field: 0 }, { $set: { field: 1 } } ); ?
[21:29:22] <ranman> proteneer: I think there are legal reasons you can't include certain things around TLS but I could be wrong
[21:42:01] <LoneSoldier728> hey is this query wrong User.findByIdAndUpdate(user, {$pullAll: {stories: storiesRemove}, $inc: {storiesCount: -storiesRemove.length}}, because I have an off by 1 error... on the total
[22:14:19] <thearchitectnlog> i need to iterate in the child document and insert the parent object id into it
[22:28:31] <ranman> proteneer: scons because of historical reasons, if you're feeling adventurous definitely check out some of the more exotic stuff in the scons file (also we support windows and needed cross platform builds)
[22:29:10] <proteneer> does that mean even if I don't need the 2.6 features I will still need an update?