Re: [twitter-dev] Streaming API track vs. Search results
Also ensure that your client is logging the raw data as received from the socket. Sometimes this will narrows an issue down to a parsing or similar error in the client. -John Kalucki http://twitter.com/jkalucki Infrastructure, Twitter Inc. On Fri, Apr 16, 2010 at 9:01 AM, Mark McBride mmcbr...@twitter.com wrote: If you can duplicate this, can you send the exact text, tweet IDs and times of the runs? Latency on the streaming API should be better than it is in search (they're both pretty fast), so having the streaming API lag search is surprising. ---Mark http://twitter.com/mccv On Fri, Apr 16, 2010 at 8:55 AM, Mad Euchre mad.ukrain...@gmail.comwrote: I wanted to test if my program is getting all the tweets it should. My simple test was track=Palin and I timed it for exactly 5 minutes. I got 3 tweets and several replies to. Then I immediately ran this: http://search.twitter.com/search.atom?q=Palin and looked for tweets in the last 5 minutes. There were 7 results for the last 5 minutes. The 3 I got from the stream matched the oldest of the 7, so there were 4 newer that the steam didn't pick up. I don't mind if there is a slight delay and the missing 4 would eventually show up. How else can I tell if I'm getting all the tweets that I'm supposed to from the stream? Thanks, Peter -- Subscription settings: http://groups.google.com/group/twitter-development-talk/subscribe?hl=en
[twitter-dev] Streaming API track vs. Search results
I wanted to test if my program is getting all the tweets it should. My simple test was track=Palin and I timed it for exactly 5 minutes. I got 3 tweets and several replies to. Then I immediately ran this: http://search.twitter.com/search.atom?q=Palin and looked for tweets in the last 5 minutes. There were 7 results for the last 5 minutes. The 3 I got from the stream matched the oldest of the 7, so there were 4 newer that the steam didn't pick up. I don't mind if there is a slight delay and the missing 4 would eventually show up. How else can I tell if I'm getting all the tweets that I'm supposed to from the stream? Thanks, Peter -- Subscription settings: http://groups.google.com/group/twitter-development-talk/subscribe?hl=en
Re: [twitter-dev] Streaming API track vs. Search results
If you can duplicate this, can you send the exact text, tweet IDs and times of the runs? Latency on the streaming API should be better than it is in search (they're both pretty fast), so having the streaming API lag search is surprising. ---Mark http://twitter.com/mccv On Fri, Apr 16, 2010 at 8:55 AM, Mad Euchre mad.ukrain...@gmail.com wrote: I wanted to test if my program is getting all the tweets it should. My simple test was track=Palin and I timed it for exactly 5 minutes. I got 3 tweets and several replies to. Then I immediately ran this: http://search.twitter.com/search.atom?q=Palin and looked for tweets in the last 5 minutes. There were 7 results for the last 5 minutes. The 3 I got from the stream matched the oldest of the 7, so there were 4 newer that the steam didn't pick up. I don't mind if there is a slight delay and the missing 4 would eventually show up. How else can I tell if I'm getting all the tweets that I'm supposed to from the stream? Thanks, Peter -- Subscription settings: http://groups.google.com/group/twitter-development-talk/subscribe?hl=en