How to connect the twitter api connector with the loop start

Is it possible to connect the twitter api connector with a loop start node ?

Yes. Just connect them via the Flow Variable Ports (select Show Flow Variable Ports in the context menu). If you want to iterate over a table of credentials, you might want to use the Table Row To Variables Loop Start Node, which has a "direct" flow variable output port. But any Node (except for Meta Nodes) has these hidden ones.

U mean try(variable ports) ?; cant connect it 

sorry i connect it but now ? it need to start over and over again, how i create the workflow that it saves the data and keep going until two weeks


No, I mean just variable ports. Every Node has variable ports, just like the Table Row To Variables Loop Start Node and the Try (Variable Ports) Node. But most of these are hidden and can be made visible with the context menu entry I mentioned. See here for details about flow variables or do a search on this site for more information.

Saving: Many possibilities, but you mentioned csv in the other thread, so I recommend the CSV Writer, set to Append.

A loop alone is able do "start over and over again". But if you want the Workflow to wait in between iterations, you might want to consider the tips I gave in the other thread.


I need, that the workflow runs two weeks to save tweets with keywords. 


My problem is more, that i cant connect the table row, i mean i connect them, but after that, all my nodes are red and dont run

That could be caused by missing configurations, a preceeding Node that has to be executed first, or a number of other reasons. Most of the time, the log should help you narrow down the error.

So: What does the log say? (Also, anyone else got a certain song about foxes stuck in his head now? ;) )

The configuration i iinsert a default string, integer and double, for missing values u need to insert the dafault


and all nodes are red

the log says nothing cant start ^^

I somehow suspect your first Node(s) has (/have) non-connected in-ports, which most Nodes can't cope with. (If yes, maybe put a Table Creator there.) But actually, if you're hung up on problems like this, simplifying to the point of ridiculousness often helps. Also, experimenting. You sound like you're not especially familiar with the whole system yet (no offense) and especially loops (like all flow control Nodes) can be a bit tricky to work with, so maybe try working on some other parts of your workflow first where you don't need them, to familiarize yourself with it.

I did it, everything works, but now.


twitter kicks me out, and i had to connect again in twitter, do you have a node which will connect me automaticly aufter 15minutes


First: congratulations!

The way thing like this are meant to be handled in Knime is via the scheduling functionality of the server version. But workflows can be executed in a batch mode, too, so a very simple batch- or shell-script or anything else your system has to offer should suffice, too.

If you really, really, really want an internal approach, and don't care about ressources at all: I had a similar problem a short while ago and couldn't find any suitable Node, so I hacked my way around it. The following code in a Java Snippet will wait for about 15 Minutes and then return the input table unmodified:

try {
  Thread.sleep(15 * 60 * 1000 / ROWCOUNT);
} catch (InterruptedException e) {

There are several drawbacks though:

  • The Node waits about 15 minutes in total, but that time is divided into equally sized chunks for every input row. The Workflow downstream will not be executed in between, but if your input table is large, that could be very busy waiting. So it's best to use it somewhere where your table will be small. But never empty, as
  • The Node will not wait on empty tables.
  • I have not tested this approach for waiting times of more than several seconds, and not over a period of more than a few hours total.
  • This is a hack. Knime may still react completly unexpected, like interrupt the waiting, run into memory issues, or freeze.

In fact, 15 minutes is a looong time in terms of computer cycles and ressources, so I still advise you to try a more external approach like a batch script.

The part is, i need to save tweets until two weeks, for my project, i need so much tweets to use data mining on them.  I have a server where i run this programm, but the problem is u cant connect a java snippet with the twitter api connector -.-

can you look at the picture please´?

I don't see anything wrong with that, apart from a missing second loop end. That said, I don't know why you would need two loops, but hey, it's your flow. Also, just in case you're letting yourself get confused: the time generator is a nice way to create dummy input, but you could substitute it for anything else with an output.

So: it's almost ready to let it run! :)

i thought with the generator , i can let it run more than one run ?

if i let it start, it goes in the 15minutes break, and dont come out -.-, but it have to use it if it kicks me out, but i cant connect the snippet with the twitter api connector

Like I said, this is kind of a crude hack, so anything might be going on here...

it should be like this, but thats not working.



But how you would create a workflow which runs from itself more than onetime marlin ? i mean in knime ?

Probably a Java Edit Variable node would better fit for this use case.