How to read csv stored in remote repository on RM AI HUB Server

hervedelhalle

Hello everybody, 

I am new to Rapid Miner Server.
I am not able to use Read CSV operator with a file stored on the server, in remote repository. 

I have tried to write an ExampleSet from Rapid Miner Studio to the Server. 
It works if I use the option "resolve relative to process directory" but it dit not if I do not use this option. 

Afterwards, I have tried to read the ExampleSet either with RM Studio or RM Server.
It works if I use the option "resolve relative to process directory" but it dit not if I do not use this option. 

In the Read CSV operator, there is no option ""resolve relative to process directory"" and I am not able to configure correctly the path in order to read the file store in the remote repository. 

Could you please explain how to correctly configure absolute path to work with file store on server remote repositories ? 

MartinLiebig

Hi,

to be sure: Is the file stored in the repository of the server or on the hard disc of the server?

Best,

Martin

hervedelhalle

Hi mschmitz,

Thank you for your reply. I do not know exactly. I upload the file from the AI HUB web browser interface.  

MartinLiebig

Hi,

i would recommend to use open file and then store in RM Studio. Than you can always use retrieve to use the file again.

Best,

Martin

hervedelhalle

Hello mschmitz,

Thank you for your reply. I am not sure I understand well your proposal.

Thus, i will reformulate it :
You propose to open the file stored on the server using the Open File Operator and then store the file locally using the Write File operator before reading it using the Read CSV Operator. Correct ?  

It this understanding is correct, is doesn't answer my request as i do not want to run my script on RM Studio but on RM Server. I am just launching it from RM Studio by now. Afterwards, il will launch it as a service from another application. Consequently, I would like to be able to read a file present on the RM server using a script running on RM Server.

I hope it clarifies my request.

Best regards

Hervé

MartinLiebig

Hi,

no, what I propose it to use Open File first. Open file creates this purple object, which is a file object. You can actually just use a Store operator to store this not locally, but on a Server. This will then have a small floppy disc icon. You can then use retrieve on your process to get it into the process (be sure to use relative paths) and push it into read CSV.

Alternatively - why don't you just store the result of read csv on the AI Hub?

~Martin

hervedelhalle

Hello mschmitz, 

Thank you for your answer. I managed to do the following operations : 

Running a script locally : 

• Open file using the Open File Operator 
• Save the file on the server using the Store Operator. 
  The file saved in such a way is effectively represented as a floppy disk icon. 
  

Running a script on the server : 

• Retrieve the file store in the locally run script and represented as a floppy disk icon. 
• Parse the content of the file using the Read CSV Operator.

Using this method, I am forced to run a script locally in order to store the file with a format compatible with the RM server (the "floppy disk icon" object) before I could open it with RM Server run. 
I would like to avoid such steps and being able to directly read file stored on the server without this "conversion" into the "floppy disk icon" object. Could you please help me to do so ? 

Best regards