Your understanding of since_id is correct. If you specify since_id=x then you should only get back tweet ids strictly greater than x. Are you using PHP 5? If so, I strongly recommend using the .json feed instead of .atom and using json_decode() to parse the data.
-Chad On Thu, May 7, 2009 at 5:37 PM, Joel Hughes <joelhug...@gmail.com> wrote: > > Hi all, > I have looked at the API and searched this group but can't find the > answer. > > I've created some code which is doing a search and doing some analysis > on the results and then poking the results into a database. Simple > enough. > > This code then runs a few minutes later. > > I obviously don't want to process the same tweets again so I'm > assuming that I need to use the "since_id" to help me out here and to > filter out previously seen tweets. > > With simplexml in php I get this kinda XML back > <id>tag:search.twitter.com,2005:1730670929</id> > > ..so at the end of my process run I'm parsing out that number on the > end as that seems like the status_id to me > > I then store that in the database and to feed it back into the process > when it NEXT runs > > .............the issue is though > > If I run the process a couple of times on a row I can still see the > same tweets coming back - this foxes me as I would have thought that > by using status_id in this way I was effectively creating some sort of > "sliding window" where I could NOT see tweets older than this previous > id. > > Is this assumption right (and, therefore, I'm looking at a code issue > here) or have I misunderstood the nature of status_id? > > Thanks for any help/thoughts you can provide > > Joel > >