couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mikeal Rogers <mikeal.rog...@gmail.com>
Subject Re: endkey doesn't obey limit parameter and always returns entries starting with offset 0
Date Mon, 13 Sep 2010 03:54:16 GMT
That gets hard when you have the same key emitted multiple times in a view,
then you need to code up and tool moving the startkey + skip in some cases.

This isn't actually 100% solved but the easiest way to handle it is
limit+skip ATM.

-Mikeal

On Sun, Sep 12, 2010 at 10:18 AM, J Chris Anderson <jchris@apache.org>wrote:

>
> On Sep 12, 2010, at 9:39 AM, Mikeal Rogers wrote:
>
> > For pagination you can use the skip parameter instead of trying to change
> > the startkey for each pagination.
> >
> > http://wiki.apache.org/couchdb/HTTP_view_API#Querying_Options
> >
> > ?startkey="foo"&limit=10
> > ?startkey="foo"&limit=10&skip=10
> > ?startkey="foo"&limit=10&skip=20
> >
>
> You can but you shouldn't.
>
> The right way to do it is to use startkey and limit, as documented here.
>
>        http://guide.couchdb.org/draft/recipes.html#fast
>
> >
> > On Sun, Sep 12, 2010 at 12:06 AM, Henrik Skupin <hskupin@gmail.com>
> wrote:
> >
> >> Hi,
> >>
> >> For my current Couch application I'm using pagination to show only 50
> >> results per page. To navigate back and forward the endkey vs. startkey
> >> parameters are used. While startkey in combination with limit is working
> >> fine, endkey doesn't return results for [offset(endkey)-limit,
> >> offset(endkey)] but for [0, limit]. This makes pagination impossible for
> me
> >> and looks like to be a bug. An example you can find here:
> >>
> >> http://mozmill.hskupin.info/general/reports
> >>
> >> After opening the page click 'Next' twice and check the URL of the
> >> 'Previous' link before clicking on it. The endkey parameter will not be
> >> obeyed and a click on that link causes the first page to be opened.
> >>
> >> The code can be found at: http://github.com/whimboo/mozmill-dashboard
> >>
> >> Is the above problem a known broken behavior and worth filing a bug?
> >>
> >> Thanks,
> >> Henrik
> >>
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message