I'm having trouble with one of my recipe. A simple one. I just filter some rows and drop some of useless columns and everything looks great on the visualization. On my sample of 10000 rows (first records), I loose 374 records.
My problem is, when I run this recipe, the ouput have 0 rows. I try re-run mutliple time (force rebuild, smart reconstruction), copy the recipe to have a different outpu, clear the output data, close and re-open dataiku... but same problem every time.
If I disable the "filter with formula", the output data actually have rows... But if my filter is deleting all rows, I should see that inside the recipe preparation, isn't it ? Moreover, I see rows who satisfy the filter so...
If any of you have any ideas, I would be thankfull.
Thanks for your answers.
I am using Spark engine. I try in local and... it failed, after of lot time of computing. But, during this, I drop a bunch of columns to keep the writing output as fast as possible. In addition, I see warning during the computation, a mismatch between the detected type and some rows. I change this in the initial dataset.
And now it works with Spark engine.
Thanks fou your help. I hadn't though changing the engine to debug.