Submit your use case or success story to the 2023 edition of the Dataiku Frontrunner Awards ENTER YOUR SUBMISSION

Force build a single dataset

Force build a single dataset

Hi there,

Sometimes I have a requirement to build a single dataset, which could be on the end of a flow. DSS will not always rebuild the dataset if it is already populated, even if I know this data is out of date, for this reason I use a force refresh. Is is possible to do a force refresh on a single dataset, rather than having a force refresh also refresh all dependencies?


0 Kudos
2 Replies


Just a thought.  I do the following to control the scope of my "forced rebuilds".

If you go into the advanced setup of the dataset just before the recipie you want to run. There set the rebuild behavior to explicit or write protect.  Then you can use the recursive force rebuild and it will not rebuild any of the datasets prior to the dataset that has been set to explicit or write protect.

Hope that helps.

P.S. One of the things I've also learned is the flow counts and update dates do not automatically update in the project flow page in all cases.  I often have to refresh the flow page web page [Command][Shift]R/[control][shift]R in order to see the actual count and last re-build date of a dataset.  This makes me some times think that my dataset has not been re-built.

0 Kudos

Hi @ben_p,

Just to confirm that you are seeing datasets not get rebuilt if you choose the "Build only this dataset" option for the Build mode?

I don't think we have experienced this. At least I'm not aware of it. We mostly are building SQL datasets and certainly "Build required datasets" won't rebuild SQL datasets when underlying data is changed (understandable as this is not as easy to determine as with files - although it'd be awesome if it was smart enough to only rebuild datasets that needed it).