Log file Viewer
#mongodb logs for Tuesday the 19th of November, 2013
[00:02:51] <aodhol> essentially i have two objects lets say, and they are both the same except they have different values for a particular array node e.g. "obj1": {"someNode":"someValue", "someList":[1]}
[00:03:53] <aodhol> but what i'd like is for obj1 and obj2 to be merged such that it would end up being: {"someNode":"someValue", "someList":[1,2]}
[00:05:39] <testing22> do you need to insert obj2 before you combine obj1 and obj2? or could you essentially add '2' to someList of obj1 in that step?
[00:06:25] <aodhol> testing22: i suppose i could run through my data before persisting it and merge manually, but i thought that inserting / updating in Mongo would do the trick
[00:13:08] <SamHad> when i type mongod i get this message all output going to: /usr/local/var/log/mongodb/mongo.log
[00:13:44] <testing22> mongod default releases to a background process with output redirected to that file
[00:13:46] <aodhol> joannac: testing22 note the dates array, there may be multiple objects with all details the same except for different dates
[00:15:40] <aodhol> joannac: testing22 so i'd like the different dates to be added to the array of the obect already existing in the DB
[00:18:02] <joannac> aodhol: erm, so every time you add a date, you want that date to be added to the object in the date document?
[00:18:45] <aodhol> joannac: everytime I add an object that contains the same details excepting the date, i'd like that date to be added to the array of the existing object
[04:10:08] <rmsy> I have a field that is an array of hashes, and I want to sort by the number of hashes in the array that have a specified value for a specified key
[04:12:43] <jkitchen> you're basically defining a custom sort function at that point, and it doesn't look like mongodb supports that
[05:15:05] <testing22> how can i design this aggregation/reporting schema to be multitimezone friendly? http://docs.mongodb.org/ecosystem/use-cases/pre-aggregated-reports/
[09:40:02] <Zelest> is it possible to have a sparse unique index with a date field as the condition?
[10:49:36] <Aartsie> Derick: I was wondering if the PHP MongoDB driver a array with Key => value will save it as a object ?
[10:49:47] <oxsav_> hello.. I'm trying to use the simple api rest of mongodb. I'm trying to do a simple GET to the server.. I do it with success but returns me this: Resource interpreted as Script but transferred with MIME type text/plain
[10:57:20] <oxsav_> I have an application only html and javascript that have a button that only executes something like $.getJSON
[12:25:39] <_boot> Hi, is there any way to 'create' a Date object in an aggregation pipeline? I need to snap the dates to year and month but I'd like to be able to keep them as actual dates
[12:27:42] <cheeser> like this? http://www.kamsky.org/1/post/2013/11/yet-another-way-to-group-by-dates-in-mongodb.html
[12:41:03] <dccc> is it possible to create a unique compound index with a boolean field? i am trying to restrict only one copy that has field1 and is_active.
[12:49:04] <dccc> i have index( { field1: 1, is_active: 1 }, { unique: true } ) but it i don't get errors when saving another copy with the same field1 reference and is_active: true
[13:13:32] <hyperboreean> I'm getting a "cursor didn't exist on server, possible restart or timeout?" error when I do a collection.find. I tried all kind of things, lower the batch_size, no batch_size, increasing it
[13:28:52] <hyperboreean> cheeser: have no idea ... I'm going to add some timing around that, though really ? it's just a fd.write(id), it shouldn't take that long, right ?
[13:33:34] <hyperboreean> the thing is that it always gets stuck after the same number of documents ~460k out of 1.9M
[13:37:48] <hyperboreean> yeah, it has nothing to do with the writes ... the average for the first 100k was below centiseconds
[13:43:18] <redondo> I'm trying to install mongodb in debian wheezy as in http://docs.mongodb.org/manual/tutorial/install-mongodb-on-debian/. And the process hangs on getting the public key (step 1).
[15:34:28] <Thijsc> Hi all. I'm trying to help a client recover a collection that was accidentally emptied. It then turned out that there were actually no backups being made. Very stupid, I know.
[15:34:51] <Thijsc> Is there any know way to get deleted records back from a collection based on the data files and the journal?
[15:36:26] <Thijsc> I just have the data files, I tried to export them with mongodump but that resulted in a collection in the same state
[15:51:13] <Tomasso> I try to create a text index, and the largest key i access is 51 characters long.. but I get "ns name too long, max size is 128"
[15:52:54] <ichilton> By default, will changing a single mongodb server to a master & slave slow down the master? Will it wait for slaves to finish writing before returning, or is that off by default?
[15:54:12] <Tomasso> this is my command: db.mycollection.com.ar.stores_v1.ensureIndex({'store.info.categories.category':'text','store.info.categories.products.product. name':'text','store.info.categories.products.product.description':'text', 'store.info.companyname':'text'})
[15:55:29] <ichilton> Am going to go replica sets eventually, but just want to add a slave as a backup at the moment.
[15:56:14] <cheeser> ichilton: i dunno about m/s but with replSets, it would not necessarily wait fo writes to the secondaries depending on your write concern
[16:13:08] <oxsav_> but I'm trying to use the simple rest API of mongodb.. trying only to do GET operations
[16:14:00] <oxsav_> Giving me a strange error.. It seems that the request arrives to server and the browser receives the response but it's not converted to a right type
[16:16:39] <Nodex> and you probaly shouldn't put readonly data available on the web for all to see, it's not secure
[17:45:43] <Tomasso> is there some way to query a text index using find, and not the runCommand ? I need to run the query from ruby code, and runCommand is not available as a method in the ruby driver
[18:48:41] <justanotherday> I noticed that Mongodb supports embedding dosuments, can this be used to replace the JOIN commmand
[18:55:39] <TheComrade> Has anyone done anything like write a script that turns some sort of mongo query into command line parameters? So you can do something like: "cmd $(mongoquery ...)" that gets expanded? Or is this a weird approach because the query would be ungainly to use in this fashion?
[18:58:53] <testing22> how can i design this aggregation/reporting schema to be multitimezone friendly? http://docs.mongodb.org/ecosystem/use-cases/pre-aggregated-reports/ i have two thoughts: hourly aggregation only, this would cause way too much segmentation of data. or aggregating in to different timezone collections resulting in redundant data, lots and lots of it.
[19:31:22] <testing22> justanotherday: embedded documents are not like joining two documents together as they are embedded and there exists no virtual relationship to another document. sql-like joins are done with database references (DBRef) but there are limitations using this. instead of thinking in traditional rdbms rows and columns, think of a "schemaless" document such as JSON and being able to store/index/query these documents. a good example is inste
[19:31:22] <testing22> of having a (o2m) user and addresses table, let's just embed addresses in to each user
[19:32:49] <Astraea5> Hi. I've never used to Aggregate function, and I was wondering if someone could help me with a query.
[19:33:41] <Astraea5> I haven't got a pipeline yet. I'm still not sure I'm taking the right approach.
[19:34:31] <Astraea5> What I've got is a collection of documents, each with a category field. I want to count each category (got that working with $group) but also return the first 3 documents in each category.
[19:39:58] <TheComrade> simple question: What is the common way for storing auth information on a client? I want to do like: mongo query.js without having to pass auth information each time.
[19:47:32] <Goopyo> Do I pass a list of ops instead of a document? http://api.mongodb.org/python/current/api/pymongo/collection.html#pymongo.collection.Collection.update
[19:48:49] <testing22> the second argument contains the two pushes, there is no third argument in that ex.
[19:49:30] <Goopyo> oh I see then how are you putting that in a hash considering the keys $push and $push are the same?
[19:51:47] <Goopyo> saw that here too http://stackoverflow.com/questions/9823140/multiple-mongo-update-operator-in-a-single-statement
[20:14:16] <jaytaph> inside a cursor i'm doing a forEach(), how do i get the first item from an array inside each document? e.myarray.0.id doesn't not even compile right?
[20:30:21] <jaytaph> x0f: yeah. tried this. It seems that myArray returns multiple bson-objects. i can jsonprint(e.myArray) without problems
[20:30:40] <jaytaph> and in here, i can see my .id, but i don't know how to actually retrieve this value
[20:33:44] <x0f> jaytaph, would be helpful to provide a pastie/gist with your whole query and some sample data (json printed)
[20:41:57] <jaytaph> x0f: http://pastebin.com/iYwvyiRW So, basically i need to fetch all the ancestors.0.id values, and update this into document.top_document_id
[20:43:19] <jaytaph> i'm trying this with document.find().foreach() which i used before, but never tried to fetch ancestors.0.id so i'm figuring out how to retrieve this.. or maybe i must update all those documents (1.5M of them) another way :/
[20:49:43] <jaytaph> x0f: might have found something: element = e.ancestors.pop(); print(element.id); that works..
[20:50:36] <x0f> jaytaph, you are sure you can't just use the bracket accessor? array[index] to get the first element?
[20:55:15] <jaytaph> actually.. pop() doesn't work either.. it's destructive on the ancestors (and obviously, i'm popping the last item, not the first)
[20:59:45] <joshua> Can someone tell me what format this json date is in from the mongoexport dump "$date" : 1384766745205
[21:02:50] <joshua> Our app keeps saying 2013-03-10T02:31:35.000 (without the usual Z) but I don't know where its pulling it from
[21:05:22] <jaytaph> x0f: it seems to be working now with [0]. Main issue that the FIRST (and only) document, didn't had any ancestors in them :(
[22:16:45] <kfox1111> map-reduce question. If you incremental map reduce into an existing collection, does it process every item in the resulting collection, or only if there is more then one item per key?
[23:13:53] <kfox1111> map-reduce question. If you incremental map reduce into an existing collection, does it process every item in the resulting collection, or only if there is more then one item per key?
[23:52:16] <kfox1111> what happens if a reduce ends up being greater then 16mb? Does the whole job fail right away? Does it continue and then fail? Does it silently fail?