ruby - How to stop thinking "relationally" -

At work, we have recently started a project using a thoughtbase (a document oriented database). I am having difficulty learning all my relational DB knowledge.

I was wondering how many of you overcome this obstacle? How do you stop thinking relationally and begin to think initially (I apologize for making that word).

Any suggestions? Helpful hints?

Edit : If there is a difference in it, then we will notify Ruby & amp; CouchPotato to connect to database

Editing 2 : I had trouble answering. I chose one who helped me to learn the most, although I think, there is no real "right" answer, I think.

I think, after adding a couple of pages on this subject after perusing, it Depends on all types of data.

RDBMSes represent a top-down approach, where you are the database designer, which will emphasize the structure of all the data present in the database. You determine that the first, last, middle of a person Name and a home address etc. You can apply it using RDBMS. If you do not have any column for a person's homepenet, then the hard luck is the person who has a different Home Planet than the Earth; You must add a column at a later date or the data can not be stored in RDBMS. Most programmers thus do the beliefs in their apps anyway, so it's not a silent thing to believe and apply. It may be good to define things but if you need additional features in the future, then you have to add them. The relationship model assumes that a lot will not change in your data features.

MapReduce the database using something like "cloud" type, in your case CouchDB does not make the above assumption, and instead see the data of the down-up data is input into documents, Number may have characteristics. It assumes that your data, according to its very definition, can have a variety of features that could be different. It says, "I know that I have this document in the database person, in which there is a Home Planet specialty of" Etonium "and is the first name of" Lord Nibbler ", but not the last name." This model fits webpages: All webpages are a document, but the actual content of the document / tags / keys can vary by different so that you can not fit them into a rigid structure, which is higher than the DBMS pants standard This is the reason why Google has doubted the MapReduce model Roxers, because Google's data set is very diverse, it needs to be created for ambiguity by getting it So much, and may be able to use parallel processing due to data sets on large scale (which makes MapReduce trivial). Document-Database Model assumes that the properties of your data can be very different with "Response" or can be very different, and very few populated columns can be found, if any data can be stored in a relational database. When you can use RDBMS to store the data in such a way, it will actually become ugly.

So to answer your question: If you look at the database when using the database, you can not think of "relational" MapReduce paradigm because it is not actually an applicable relationship, it is a conceptual hump , You have to end now.

---

A good article I was running in comparison with that and the two databases very well contradict, which argues that the mpadus paradigm database is a technical step behind, and is inferior to RDBMSes . I have to disagree with the thesis the thesis and he will present that the database designer has to make the right choice for his position.