'There was nothing to do for this job' when running scenario

[patwatt]

What does it mean when I run a scenario and a job in the scenario returns the message "There was nothing to do for this job".

For example, my scenario contains a Teradata dataset built from a SQL query. I sometimes get the "There was nothing to do for this job" message when this dataset is to be built in the scenario run.

How does Dataiku identify that nothing needs to be done for this job? Does it use the connection to the Teradata database to identify that there is no new data in the database since last time the dataset was built?

[CatalinaS]

Hi @patwatt,

The message "There was nothing to do for this job" indicates that all the dependencies that needed to be built for the job were already built and up to date and there was essentially nothing needed to do for the job. 
​
If you are using a build step for your scenario, you can set the "Build mode" field to "Force-rebuild dataset and dependencies" to guarantee that your dependencies will be rebuilt. 

[patwatt]

I already have the "Build mode" field set to "Force-rebuild dataset and dependencies". But I still get the message "There was nothing to do for this job" for one of the scenario steps. This is concerning, as I need to build the datasets every time the scenario runs. Perhaps it is a bug? In addition, the step that gives this "nothing to do" message has datasets that have not been built for over a week. There is certainly new data at the source that needs to be accessed.

So, how does dataiku determine if a dataset is "up to date"? Is there an amount of time after which it considers a dataset is not up to date? What is that amount of time?