I wouldn't, ayhank. I would use a batch/shell script to let the operating system start Knime all 15 minutes. Or I would use something else for the data gathering phase, because Knime is an analysis tool first and foremost, not a data gathering tool.
As for aborgs suggestion, I did some quick tests and my trick doesn't seem to work with a Java Edit Variable. But maybe it does for you...
In general, I agree with Marlin here, except I would use the *cough* commercial *cough* KNIME Server :)
Alternatively, if you are really motivated to do this in a single workflow, see the attached screeshot for my approach. The problem with using a CSV writer + appending data is that you may catch the same tweet many times, which will be messy (but certainly not too difficult) to deal with later. Doing an update on an existing SQLIte database of tweets first gives you the added benefit of being able to update retweet & favorite counts.
To run the workflow you would just specify the number of iterations in the loop start node and the time gap for each iteration in the wait-for-time node. Post back and let us know how it goes.
Hi people its been a while since last post on this thread,
I am trying to do a Keyword research using KNIME. As I need a high ammount of data I will need to create a loop just as shown above on pic to do the job. But I have a question, how do I create a template database to start updating it with all the data I will collect from Twitter. I need I a database to start working with as SQLite Connector requires it. I am a little lost here.
Thanks!