From: <ni...@br...> - 2001-08-03 16:03:37
|
thanks for the anwer. i 'm using openldap (not sure if this has server side sorting), but if it does, it may be worth me looking into the jndi before using caching. boo hiss ;-) nick Michael Str=F6der wrote: >=20 > ni...@Br... wrote: > > > > Ideally, I want my search interface to return just 20 records at a ti= me, > > which are a sorted subset of a larger result set. Does the LDAP > > protocol allow this? >=20 > The right solution would require deploying two LDAP extended > controls: > - paged results (OID 1.2.840.113556.1.4.319, see RFC2696) > - Server Side Sorting of Search Results > (OID 1.2.840.113556.1.4.473, see RFC2891) >=20 > But unfortunately > 1. with python-ldap you can't currently make use of extended > controls and > 2. the server has to support both controls (which you can > check by looking at attribute supportedControl of its > RootDSE). >=20 > > I'm trying to think how else to do a multiple-page > > search interface, ala current search engines with the footer >=20 > Well, you could retrieve all search results, sort them on the client > side and cache it for the next page hits. But take care of security > considerations! >=20 > My web2ldap cheats: It simply invokes a async search and retrieves > the search results for the currently viewed page without doing any > sorting. I considered the user to give up looking through the search > results on the first page anyway. ;-) >=20 > Ciao, Michael. --=20 Nick Bower, Intranet Developer ni...@br... Brainstorm 388 - 396 Oxford St London W1N 9EH United Kingdom Tel 020 7074 7000 Mob 0790 5405 443 Fax 020 7074 7070 http://www.brainstorm.co.uk |