[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Granularity

• From: Michael Sokolov <sokolov@ifactory.com>
• To: Andrew Welch <andrew.j.welch@gmail.com>
• Date: Fri, 06 Jan 2012 12:51:34 -0500

On 1/6/2012 9:55 AM, Andrew Welch wrote:
>> Many XML databases work best when you have lots of small documents; a few
>> work well when you have one giant all-embracing document.
> The problem with one large doc is you will only get 1 hit as a result
> for multiple matches... instead of n hits for n docs that each contain
> a match.
>
> You can work around that, but in general its best if your docs
> correspond to the number of results you would expect for a given
> search.
>
>
I guess it depends on how the indexing works - you might be able to 
index at a sub-document level and search using xpath within documents; 
xslt lets you do this on the fly at least.  But it's true that most of 
the well-known xml dbs index at the document level.  I guess in 
MarkLogic they have a way to "fragment" sub-documents, but the 
recommendation generally seems to be not to use it, and to fragment your 
documents yourself.

-Mike

• References:
  • Granularity
    • From: "Len Bullard" <cbullard@hiwaay.net>
  • Re: Granularity
    • From: Michael Kay <mike@saxonica.com>
  • Re: Granularity
    • From: Andrew Welch <andrew.j.welch@gmail.com>