How to change string/text data type to double/decimal throughout a build?

[jrod31]

Hi all! 

This the following scenario on a build:

I have a dataset D1 pulling in data from a SharePoint folder with several CSV files. One of the columns in the D1 dataset has decimal numbers which contains a % symbol. Dataiku reads it as a string/text data type. 

If I change the data type to double/decimal - all the data in that column won't show. I also added a cleaning recipe to take out the % symbol and tried to change the data type to double/decimal and I still loose the data in that column. 

I have to switch everything back to String/text in the data type, so all follows through. 

Can I change the data type in the schema at the end of the build or can someone guide me on this? 

Thanks!

---

Operating system used: Windows

[Turribeach · Accepted solution]

Try using the Extract Number processor in a Prepare recipe.