A
A
alesto2011-06-06 10:49:38
SQL
alesto, 2011-06-06 10:49:38

Recommend literature that will "shift the brain" from Sql thinking to Nosql

The other day I tried to deal with mongo DB and came to the conclusion that I can’t imagine how I would store the data by transferring them from mysql to nosql. The articles I found partially displayed the functionality of this solution. But they did not answer the question “How to organize all this at the project level?”. Please advise books \ articles in Russian or English that can change sql thinking towards nosql. Preferably using mongo db or cassandra as an example. Thank you.

Answer the question

In order to leave comments, you need to log in

5 answer(s)
A
Alexander Korotaev, 2011-06-06
@aavezel

In my opinion, everything is simple here. If you can't figure out how your project data will be stored in NoSql databases, then you don't need a NoSql database in your project. NoSql is not SQL + goodies, NoSql is SQL - (minus) unnecessary functionality. Therefore, IMHO, books or articles of the content “Moving to MongoDB. For dummies" cannot be in nature.

I
INSTE, 2011-06-06
@INSTE

Map-Reduce + nosqldb with support for microtransactions (many operations performed elementarily) is a good start. But the data model will have to be maintained by hand, including building reverse indexes and so on.
But without caching and a properly built system, you will lose 100% both in speed and in memory. Nosql helps to parallelize the task as much as possible and “fill up with iron”, but this should be done not as a replacement for sql, but from scratch, without looking back at the sql past.

Y
Yeah, 2011-06-10
@Yeah

And here is a good video where a friend tells how he developed a complex system first on Postgres, and then on CouchDb:

V
VBart, 2011-06-07
@VBart

Look at the book on CouchDB, I don’t know how it can “shift thinking”, but it seemed to me that it was not bad. It's also free: guide.couchdb.org/

I
iryndin, 2014-07-09
@iryndin

On monge here: Data Models
Carefully read, make examples. Video from previous answer - cool! See him also.
The general idea is to dump the data in one heap (denormalization) and make a minimum of requests to the storage.

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question