Bibliography http://baike.baidu.com/link?url=DVJuvrT4ssbM8V1pKcoRx5H1CwRRCynC6_ Ix6ywirr414opislflppuy9hrdo4g-fr5vl9afbsvamwagpvn8g_
Book over + read blog (http://www.cnblogs.com/huangxincheng/category/355399.html) + DIY installed MongoDB wrote demo code (windows,c#)
Summarized as follows:
- The language is good and looks comfortable. Like natural language, JavaScript, json feeling >db.food.find ({"Age": find,insert,update), good, etc.
- document (record, row) = "collection (table) =" Database
The
- is generally very easy to expand, without the need to define the table structure (various column names, types) in the same way as the relational database (SQL Server), and you can insert the person {no,name,sex ...} directly.
- can also write directly to the complex type person.adress{provice,city,street,detail ...}, so that you do not have to like the previous relational database as entangled in the multi-table foreign key connection, special trouble. MongoDB this point, I like it very much. This also ensures that the joint check is very fast and simple.
- When manipulating MongoDB using the program language, the corresponding driver needs to be installed according to the language. If you use the C # language, then using the driver (https://github.com/mongodb/mongo-csharp-driver/downloads) is very simple, similar to the linq,ef feeling, do not have to dwell on the specific CRUD statements, As long as the getcollection<t>,update can. Of course it's just my first coding impression, and maybe the details are pretty complicated.
- shards, which are used to balance data, reduce data on each machine, reduce traffic, and use multiple machines to scale effectively without having to buy large machines
- primary and standby replication is used to ensure multiple backups of data, redundancy prevents accidental, but still need to manually switch master and standby changes. Replica set for automated disaster-tolerant clusters, with mechanisms such as quorum machines, replicas, active, heartbeat judgments, and automatic guarantees that one cluster must be available. Both do not need program control, transparent procedures, black boxes, no need to care. This is good for programmers-dba the
- index. Similar relational database
- aggregate function count,group. The physical storage entity for a similar relational database
- document is a binary JSON (Bjson) type, which is lightweight like JSON text and can represent all documents as strings.
- Basic Data type: number, string, date, array, inline document (subdocument, function point mentioned above 4)
In general, MongoDB is not a relational database, and it has advantages in big data processing, complex data processing and balanced allocation processing.
Official website Http://www.mongodb.org/downloads
After learning, although the technology in the development, but the basic concepts, ideas and ideas are similar. Understanding the relational db, Javascript, Json and then look at MongoDB, obviously easy, easy to understand, the development of technology has both same strain, and change, to solve the old contradictions proposed new solutions.
The learning Summary of the MongoDB authoritative guide