Sychronize Metadata with Real data

Niharika

I am building a process from in which I gives dataframe from my "Execute Python" operator which has to be used by read csv/process further used. Sometimes in sub-process after python the data is not fetched.
In Tab-->Process (I checked the option) Synchronize Meta data with real data.
But how do I ensure that everything is synced when I run the process on server.

Marco_Boeck

Hi,

Meta data is ONLY used while you design the process in the UI, i.e. to show you a pre-populated combobox of available attribute names and such. It has NO other use. Meta data has absolutely zero impact when running a process.

Regards,
Marco

hughesfleming68

The quick way to restore your metadata is to store it to the repository and then retrieve it. In your example, store it after the execute python operator and then retrieve and write your CSV file. You may have to adjust the order of your operators when your process is not continuous so be aware of that.

There is a more formal way that is outlined in the execute python documentation that involves describing your metadata in your python or R code. Take a look at the section that discusses handling metadata towards the end.

https://docs.rapidminer.com/latest/studio/operators/utility/scripting/execute_python.html

regards,

Alex 

Niharika

I went through the tutorials but I not very clear.The first answer you provided would definitely work but is there other way round to synchronize the meta data.
I have created the above process which gets data for n machines from python and creates n no of models which I store in my same subfolder of above process.But when I am running the above process on server it runs successfully but no new models are getting reflected in my repository.
Probably due to unsynced meta data.So I am not able to fix the issue.