Hi Malcolm,

> To help conserve resources, should indexing be removed once used?
>

No. It would be very inefficient. Index creation is considered as expensive
operation which should be done usually at the beginning of the deployment,


> Can indexes become corrupted?  If so, is there a method to determine the
> corruption and correct?
>

No (otherwise it's considered as a bug). Indexes are updated automatically
whenever data is updated.


> In some situations, can an index save memory?  For instances, rather than
> conducting a full scan on a very large file to return data (which would
> require loading the file into memory), can using index prevent the full file
> load?
>

Yes! That's exactly why do we need them. For example, we got:

<persons>
    <person name='Malcolm">...</person>
      ....
   <person name='Ivan'>...</person>
</persons>

and want to get <person name='Ivan'> element. In naive straightforward
implementation we have something like:

doc('persons')/persons/person[@name eq 'Ivan']

it will scan ALL <person> elements to find the last one. For example, you
have millions of persons. Sedna would need to read from disc all chain of
blocks which contains <person> nodes.

But if you have index that query looks  like:

index-scan("person-by-name", "EQ" "Ivan")

That's it. It's not direct access but it's much much better than full scan.

Sorry, I am not a dba type, so I am unaware of the problem space and
> techniques used in db development.  I reviewed the Sedna technology papers,
> but I am still clueless.
>

No problems. Do not hesitate to ask any questions.

Ivan Shcheklein,
Sedna Team
------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security 
threats, fraudulent activity, and more. Splunk takes this data and makes 
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2d-c2
_______________________________________________
Sedna-discussion mailing list
Sedna-discussion@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sedna-discussion

Reply via email to