Tag: nosql

MongoDB Testing

MongoDB is a NoSQL type of database.

Installation on Windows is rather simple just follow the prompts. Just download from here. You can follow the installation instructions from here if you so choose. It has a max document size of 16MB but you can see all base configurations from here. Once you install MongoDB you will have to start the server. Run C:\Program Files\MongoDB\Server\3.2\bin\mongod.exe. Then you can run C:\Program Files\MongoDB\Server\3.2\bin\mongo.exe to start querying the data.

MongoDB is a document db. Adhoc querying is easy if the data you are querying is within the 16MB max size. There is no tables in MongoDb they are called Collections instead. They are used to store bson documents. Which are basically json. Querying data is fast I have found. It uses the map reduce context to querying data. If you are not familiar with map reduce it’s basically just JavaScript.

I will have more later as I test more and more.

I have done some testing and can give you some basic code below.

Show All Collections:

 show collections

 

See All Databases:

 show dbs

 

Drop A Collection:

 db.COLLECTIONNAME.drop()

 

Drop All Collections:

 db.getCollectionNames().forEach(function(c) { if (c.indexOf("system.") == -1) db[c].drop(); })

 

Create A Collection:

 db.createCollection("logs", {autoIndexId: true})

 

Create An Index:

 db.logs.createIndex( { KEY: VALUE } )

 

MapReduce:

 db.COLLECTION.mapReduce(function(){
      var key = WHATEVER;
      var value = WHATEVER;

      emit(key, value);
},
function(key, values) {
      var output = {
            KEY: VALUE
      };
      return output;
}, {out: {inline: 1}})

 

Find matching element:
The .pretty() displays the json pretty.

 db.COLLECTIONNAME.find({ "KEY.SUBKEY.ARRAYINDEX" : VALUE }).pretty()

 

Find record that match in using greater and less than or equal to and display the id field associated.

 db.COLLECTIONNAME.find({ "KEY" : {$elemMatch: { $elemMatch: { $gt: 0.0005, $lte: 0.0005862095 } } } }, { id: "_id"  }).pretty()

 

Find record using IN and returning id field.

 db.COLLECTIONNAME.find({ "KEY" : { $in: [ 123,456 ] } }, { id: "_id" }).pretty()

 

You can loop through data and insert it into a new collection

 db.COLLECTIONNAME.find({ "KEY.SUBKEY.ARRAYINDEX" : 0 }).forEach(function(obj){ db.NEWCOLLECTION.insert(obj) })

 

CouchBase Testing

CouchBase is a NoSQL type of database.

Installation on Windows is rather simple just follow the prompts. Just download from here. Unfortunately at this time CouchBase on Windows wants you disable the firewall. I don’t recommend this and due to this critical issue itself do not currently recommend this until it has been fixed. Once installed it is viewable from http://localhost:8091. It has a max document size of 20MB but you can see all base configurations from here. CouchBase is a document db. It has fast importing of documents. CouchBase has views as like CouchDB. A view in CouchBase is like querying for data but not like CouchDB’s views. CouchBase still has Index’s. It’s view has a fast view rebuild which makes querying data faster than in CouchDB. It uses the map reduce context to creating views. If you are not familiar with map reduce it’s basically just JavaScript.

I will have more later as I test more and more.

I have done some testing with a view and can give you some syntax for writing one.

 function(doc, meta) {
      emit(meta.id, output);
}

CouchDB Testing

CouchDB is a NoSQL type of database. CouchBase is related but different entirely.

Installation on Windows is rather simple just follow the prompts. Just download from here. Once installed it is viewable from http://localhost:5984/_utils/. It has a max document size of 4GB but you can see all base configurations from here. CouchDB is a document db. If you come from a traditional relational setting then you might get frustrated at first because well querying documents is slow at first because you need to create a view. A view basically is kind of like an index from relational setting. Each different way you query the data you need to create a view. What get’s a little ridiculous is that as your data grows the view isn’t updated with newest data till the view is called. So at first it’s slow but each time it gets queried it gets faster. It uses the map reduce context to creating views. If you are not familiar with map reduce it’s basically just JavaScript.

I will have more later as I test more and more.

I have done some testing with a view and can give you some syntax for writing one.

 function(doc) {
      //Add any code you want
      emit(key, output);
}