Twitter.text picks up RTs as well?


#1

Based on your documentation (example here: http://dev.datasift.com/blog/csdl-optimization-techniques) I expected that the twitter.text target is only applied to original tweets that are not RTs. However, i was running a few tests and it appears that this target does pick up some RTs as well. Is this a normal behaviour? If so, how can I exclude RTs from my collections?


#2

Hi oltondev,
You’re correct in your understanding that all twitter.* targets do not include retweets. Looking at the output format of a retweet, there is no twitter.text field.
Could you post a simple interaction that appears to be a retweet?


#3

Hi,

The one below is one of these:

{“interaction”:{“schema”:{“version”:3},“source”:“HootSuite”,“author”:{“username”:“SeanDElder”,“name”:“Sean D. Elder”,“id”:175484223,“avatar”:“http://pbs.twimg.com/profile_images/378800000123728516/048e822d86db6d7688e06e9461a21345_normal.jpeg",“link”:“http://twitter.com/SeanDElder”,“language”:“en”},“type”:“twitter”,“created_at”:"Wed, 19 Mar 2014 13:50:27 +0000”,“received_at”:1395237027.1951,“content”:"RT @gamespot “Sony Unveils PS4 Virtual Reality Initiative “Project Morpheus” - GS News update. http://t.co/HAx47oAADn"...",“id”:“1e3af6d6d391a380e0740345211d9460”,“link”:“http://twitter.com/SeanDElder/status/446282555604889600”,“mentions”:[“gamespot”],“mention_ids”:[7157132]},“salience”:{“content”:{“sentiment”:4}},“twitter”:{“created_at”:"Wed, 19 Mar 2014 13:50:27 +0000”,“filter_level”:“medium”,“id”:“446282555604889600”,“lang”:“en”,“media”:[{“id”:446156633182580740,“id_str”:“446156633182580737”,“media_url”:"http://pbs.twimg.com/media/BjERGBMIMAEwDor.jpg",“media_url_https”:“https://pbs.twimg.com/media/BjERGBMIMAEwDor.jpg”,“url”:“http://t.co/HAx47oAADn”,“display_url”:“pic.twitter.com/HAx47oAADn”,“expanded_url”:“http://twitter.com/gamespot/status/446156633425854464/photo/1”,“type”:“photo”,“sizes”:{“thumb”:{“w”:150,“h”:150,“resize”:“crop”},“small”:{“w”:340,“h”:191,“resize”:“fit”},“medium”:{“w”:599,“h”:337,“resize”:“fit”},“large”:{“w”:1024,“h”:576,“resize”:“fit”}},“source_status_id”:446156633425854460,“source_status_id_str”:“446156633425854464”}],“mention_ids”:[7157132],“mentions”:[“gamespot”],“source”:“HootSuite”,“text”:"RT @gamespot “Sony Unveils PS4 Virtual Reality Initiative “Project Morpheus” - GS News update. http://t.co/HAx47oAADn"...",“user”:{“name”:"Sean D. Elder”,“url”:“http://www.darksunfilm.com”,“description”:“I am a child of Adonai, a husband & father, cinematographer, musician, gamer, and story-teller. I am getting back into shooting again, no matter the format !!!”,“location”:“Houston, Texas 77077”,“statuses_count”:7404,“followers_count”:747,“friends_count”:2003,“screen_name”:“SeanDElder”,“profile_image_url”:“http://pbs.twimg.com/profile_images/378800000123728516/048e822d86db6d7688e06e9461a21345_normal.jpeg",“profile_image_url_https”:“https://pbs.twimg.com/profile_images/378800000123728516/048e822d86db6d7688e06e9461a21345_normal.jpeg”,“lang”:“en”,“time_zone”:"Central Time (US & Canada)”,“utc_offset”:-18000,“listed_count”:24,“id”:175484223,“id_str”:“175484223”,“geo_enabled”:false,“verified”:false,“favourites_count”:399,“created_at”:“Fri, 06 Aug 2010 18:51:31 +0000”}}}


#4

Hi,
Thanks for posting a sample JSON interaction.
Looking at the JSON itself shows no retweet field, thus it will be filtered as a tweet by DataSift.
While the content of the message is clearly a copy, http://twitter.com/SeanDElder/status/446282555604889600 , it was not retweeted within Twitter in the conventional sense.
You can choose to filter out tweets with “RT” in the text using case sensitivity. Something like:

AND NOT twitter.text contains [case(true)] "RT"

Remember to use this in combination with other filters.
Hope this helps!


#5

Thanks for your help - this is useful to know!