jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Lukas Kahwe Smith <...@pooteeweet.org>
Subject Re: SQL2 (via davex)
Date Wed, 10 Aug 2011 10:14:03 GMT

On 10.08.2011, at 12:04, Jukka Zitting wrote:

> Hi,
> 
> On Wed, Aug 10, 2011 at 11:46 AM, Lukas Kahwe Smith <mls@pooteeweet.org> wrote:
>> I explained the use case in my original post. What I need to do is to
>> get the last X days within /issues sorted decending by date
> 
> I was rather interested in the reason why you need the list of last X
> days, i.e. what value does some client of your application get from
> this information? A more typical query for an issue tracker system
> (which I assume you're building) would be to ask for the last X
> issues.

Ah. the use case is allowing users to select one of the last X newspaper issues available.
issue means newspaper here :)

> Anyway, assuming the date nodes really are needed for more than just
> organizing the underlying issue nodes, I'd recommend labeling them
> with their own node types and adding relevant information like the
> date as properties of those nodes. Parsing information from the path
> (like in this case the date) can lead to unnecessarily complex code
> like we're seeing here.
> 
> With such a node type, you should be able to get the last X days with
> the following query and a limit of just X results:
> 
>    SELECT * FROM [my:day] WHERE ISDESCENDANTNODE('/issues') ORDER BY date DESC


Well I would still need to "denormalize" the date which is already part of the path, but I
would spare myself having to also store the depth if I would use a node type.
But I wasnt aware I could use  ISDESCENDANTNODE() like this, thx!

So to summarize I would still like to see PATH() (so that I can sort on that) and DEPTH()
(so that for this use case I would not need a custom node type) implemented :)

regards,
Lukas Kahwe Smith
mls@pooteeweet.org




Mime
View raw message