couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Zdravko Gligic <zgli...@gmail.com>
Subject Re: few doubts
Date Tue, 05 Jul 2011 06:43:27 GMT
While I am barely scratching the surface with CoudhDB, I would go out
on the limb and suggest that it is next to impossible to pre-determine
how happy you will end up being with CouchDB as a general replacement
to MySql or any other relational database because they are two
different things.  Therefore, you will more likely need to start by
not comparing them and instead more precisely focusing on what it is
that you are trying to achieve with CouchDB. Since you are familiar
with MySql, what is it about it that brought you looking for an
alternative? Those are perhaps the most important questions because
whatever the answers are, those answers would most likely help to
explain at least some of the differences as well as some of the
reasons why they can not be that closely compared.   Being new to
CouchDB, it might be quite useful to also consider whether you intend
to host it yourself or (like me) you plan to use www.IrisCouch.com or
www.Cloudant.com.

In any case, consider what would happen with MySql if it had no
indices and if it had to scan all of the records in queried tables.
Well, much the same thing happens with CouchDB without any views.
However, if you add some indices onto main MySql tables and columns
then you are able to more dynamically and more efficiently query those
table columns in variety of non pre-determined combinations.  This is
not the same in CouchDB.

CouchDB being "schemaless" means that you can indeed create documents
of various "shapes".  However, you still have to have a very "strongly
typed schema" - both in your head and in your design docs code, in
order to be able to get anything useful out of CouchDB, beyond
directly accessing individual documents by their build-in "_id".
Specifically, it means that you will need to pre-determine each and
every index/sort combination by which you need to access those
documents.  Once that is done, then CouchDB seems to be both efficient
and scalable in what it does.

Finally, before you consider any of this, let's wait for our technical
friends to either correct and/or to reduce to nonsense whatever I just
wrote. ;)

HTH,
teslan

Mime
View raw message