Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

Tricky paths (Windows)

Hi all,

I'm just starting out on RapidMiner but am having a great time learning. ;D I wanted to get your input on a tricky problem. I'll be sharing a process with a number of people (more than 10) so I wanted to send a full repository that includes both process, data files, and a few plain text lists of word exclusions. It's important that relative paths are used as the paths will be different for each user's computer if they store in the typical location (anything under a Windows username directory, e.g. C:\Users\[USER]\blahblah).

For the life of me, I can't figure out how to input a relative path into the Read XLS and Open File operators. I have reviewed documentation both here: http://docs.rapidminer.com/studio/operators/repository_access/retrieve.html and https://rapid-i.com/rapidforum/index.php?topic=2350.0

Specifically, if I try to use "Special_words_cleaning.txt" as the filename parameter the error indicates that the path resolves to "C:\Users\02PUR\Special_words_cleaning.txt" even though my repository is located at "C:\Users\02PUR\.RapidMiner5\repositories\Hackathon".

The behavior doesn't seem to match what was described in the previously linked forum response. Has anyone good advice here? ???

Thanks!

Find more posts tagged with

AI Studio

Accepted answers

All comments

MartinLiebig

Hi,

you are referring to a really old post. Sebastian posted something regarding RM 4. I would not really trust in that anymore. A lot changed.

The documentation is for the Retrieve operator. The retrieve operator is used for the repository, not for your hard disc.

I am not sure how relative paths should work on hard disc. Especially i am unsure to which path it should be relative? Your Repo? Your home folder? Your .Rapidminer folder? And i am not sure whats realized in our software. Maybe it is possible to call some windows equivalent to $HOME. I will take with Marius about this - but maybe some other user/marco has an idea on that.

By the way: Do you have accsess to a Rapidminer Server (rapidanalytics)? It would really help you if you work together with 10 persons or so. I would recommend the Version 6 there. If you are using this on a public university/colleage,you might want to have a look on our academics program.

Cheers,

Martin

seanpuri

Thanks Martin, I was thinking of addressing data relative to the repository folder so that others can just add the repository (that includes the process and data) and go.

Right now, it seems the default path is the root of my User folder rather than the repository folder (e.g. if "./" worked).

I will look into the server and Version 6 as well for future events (this one is tomorrow so there is less time).

Regards,
Sean

Marco_Boeck

Hi,

there is an important distinction between the repository and your regular harddisk. The repository is a place which Studio uses internally to store and retrieve data from. It is irrelevant where the data is actually stored (Harddisk, Database, etc). Your harddisk is - well - your regular harddisk you can access with something like the Windows Explorer. Relative paths for the harddisk don't make sense.
You can however use relative paths when accessing a repository. A process is stored in a certain location of your repository and as such you can use relative paths ("Retrieve" operator)

There is however also a way to use files with relative paths: You can store them as binary objects in your repository. You can then use an "Open File" operator to access the repository by switching the 'resource type' parameter from 'file' to 'repository blob entry'. This file can then be sent into the input port of a "Read Excel" operator. In other words, as long as you store everything (even Excel files or whatnot) in your repository, you can use relative paths to access these elements again.

Regards,
Marco