Stress-testing SQLite -- millions records? Use Valentina DB

[stephen barncard]

Yes, I've been waiting for Ruslan to chime in here.   Valentina has been the
*elephant in the room* in this discussion and I find it slightly odd that
Richard (no newbie in the Rev world) hadn't considered this product for his
project.

If I were starting a new db project right now and wasn't forced into mySQL
by the client, I would take a serious look at the Valentina  ADK.

Right now they are offering the beta of Valentina Studio Pro for free ( and
there's a free Valentina Linux server for non-commercial use - Richmond?)

geesh, I just talked myself into finally trying this product myself.

I don't see any other db company bending over backward to serve Rev/Livecode
users.   And we even have Ruslan on the list here....


On 29 October 2010 23:28, Ruslan Zasukhin
<ruslan_zasukhin at valentina-db.com>wrote:

> On 10/30/10 3:17 AM, "Richard Gaskin" <ambassador at fourthworld.com> wrote:
>
> Hi Richard,
>
> > I have a need coming up for a data store that can robustly handle at
> > least a million records, ideally up to five million, where each record
> > may be as large as 5k.
> >
> > I don't need relationality, so for me SQLite is an option but only an
> > option; I'm happy to consider other options as well. (Yes, it has to be
> > SQLite rather than MySQL, because it needs to work embedded with a
> > commercial application).
>
> Valentina DB is faster 100 times of SqlLite, mySQL
>
> And can be perfectly embedded into commercial application because it is
> royalty free.
>
>
> > Have any of you done stress testing on SQLite to that degree?
> >
> > I've tried finding even anecdotal data on the web for SQLite limits, and
> > while I can find citations of theoretical limits I haven't come across
> > real-world usage stories of data sets that large.
> >
> > Should I be confident in SQLite as a storage solution for that?  Should
> > I be scared?
>
> SCARED :)
>
> > Thanks in advance for sharing your experiences with large data sets if
> > SQLite.
>
> You need Valentina DB.
>
> Okay you need 5K for each record.
> How many fields?
>
> Let me remind that Valentina has columnar format.
> This is huge advantage.
>
> Also Valentina can give you not only SQL way but NON-SQL way,
> Which can be additionally 10-20 times faster!
>
> ----------
> I can tell you store, that Valentina was tested for AOL Europe by their dev
> team. Against Berkly,  mySQL, postgre, and other dbs. SqlLite even was not
> in game of course.
>
> Task was so simple. Table has 2 fields {URL, PictureBannerAd }
>
> So when somebody ask for a WEB page, it needs find banner to be shown.
>
> As they told, e.g. Berkly have give 100 faults per time (min our hour I not
> remember now).  Other dbs also. Fault means that banner was not found  by
> DB
> in time less of timeout.
>
> Valentina have give them zero faults.
>
>
> Let me underline this very important feature NON-SQL-ness of Valentina.
> As well as very powerful SQL.
>
> Today is very modern stream talk about how SQL DBs are bad, and how cool
> are
> NON-SQL with Key-Value.  Guys, be happy, Valentina is perfect for both
> tasks. :-)
>
> If talk about details, in V4REV API (and most others Valentina ADKs) you
> can
> use not SQL way to do searches and sortngs using
>   VField_FindValue()
>   VField_FindRange()
>   VField_FindLike()
>
> And other similar search methods.  They are really FASTEST POSIBLE way.
>
>
> --
> Best regards,
>
> Ruslan Zasukhin
> VP Engineering and New Technology
> Paradigma Software, Inc
>
> Valentina - Joining Worlds of Information
> http://www.paradigmasoft.com
>
> [I feel the need: the need for speed]
>
>
> _______________________________________________
> use-revolution mailing list
> use-revolution at lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your
> subscription preferences:
> http://lists.runrev.com/mailman/listinfo/use-revolution
>



-- 



Stephen Barncard
San Francisco Ca. USA

more about sqb  <http://www.google.com/profiles/sbarncar>