Challenge #273: Filling In The Blanks

[Spalders]

Révélation

[MatthewBr]

Nice challenge - Never used Generate Rows before now. 

[CandiceDB]

My solution:

Révélation

[TonyA]

This one is more of a "stream of consciousness" workflow. I didn't plan it out, just started throwing tools on the canvas. Certainly not the recommended way of building out a workflow 🙂. I'm sure it could be tightened up quite a bit.

Révélation

[JeffF]

Révélation

[AkimasaKajitani]

My solution.

Révélation

[kelly_gilbert]

My solution is pretty similar to many others:

Révélation

Since this was a pretty quick one, I decided to experiment with some slightly different configurations, to see if there were performance differences:

Révélation

I wondered how big of a performance hit I was taking, since I opted for the convenience of the Data Cleansing tool. I tried replacing it with a Formula tool to replace the nulls with zeroes.
Next, I tried moving the Formula tool (to fill in the zeroes for the missing records) to the left output of the join only.

Results:

Révélation

I randomly generated a larger test dataset with 2,600 unique client IDs and 636K input records (with the workflow adding 300K records to fill in the blanks), and then ran each configuration 30 times on the same dataset.

Even at that much larger dataset size, the workflow ran in < 5 sec, and Data Cleansing tool added less than 1 sec to the average run time. I was expecting a larger difference. It would be interesting to see how that plays out with an even larger dataset.

Moving the formula tool didn't make a measurable difference in run time, although the measurement isn't very precise.

[DaisukeTsuchiya]

Révélation

 

[kelvin_law1]

Here is my solution:

Révélation

[mike_w]

don't take this as an excuse to do any non-essential online shopping! 

Révélation