Read CSV doesn't work

dajoedic

Hello,

I didn't work for a long time with Rapit Miner and I have a problem with reading csv files.
I use the "Read CSV" operator, select my csv file with the "Import Configuration Wizard", follow the steps and want to do some kind of clustering afterwards. When I run the prozess there is always the Error Message: "No input file was defined. No file input defined: depending on the source type configuration of your operator, you have to either connect the file input port or define the parameters which specifies the file to be read."

Reading Excel files work fine.

I'm using RapidMiner 5.3.013.

For more information please contact me.

Thankls for your help.

David

homburg

Hi dajoedic,

it looks like there is no file associated to the "Read CSV" operator. Did you use the import wizard directly from the parameter tab of the operator? It would be great if you could post the xml code of your process here.

Cheers,
   Helge

corkie

with a new process locate the read CSV operator
add it to the main process window
wire the op port to res
click on the wizard
locate the csv file
click next
in the file reading confirm encoding = UTF-8
in the comma separation note the default = ';'
you need to set this to ','
NEXT
NEXT
FINISH
data should be there now.

if its not the case then as per the previous response you will need to send the XML code for analysis

hi_joshi

Unfortunately, the steps provided to address this issue are not very helpful. They seem to assume a lot of things. For example, how does anyone who is using RapidMiner for the first time to get a feel of it, know how to generate an XML file for the process? 

 

Besides, there are several other issues I am noticing right now with the overall UI itself. I am unable to find the "Data" option which I saw right after installation which I could use to play with IRIS data. I have restarted RapidMiner so many times after that and was never to get that option again. 

 

I read the reviews about RapidMiner and they all seem to be speaking very high of it. I was very hopeful to see a good and intuitive tool when I read those reviews. But I am not really pleased with what I have seen so far. Maybe, it might be something at my end, but looks like a lot needs to be done, especially testing of the tool, at the RapidMiner end as well. 

 

Hope someone will take notice of this feedback. 

 

Manoj Joshi

 

hi_joshi

After some failed attempts, I finally got make it work. I think the tool has a lot of potential and I want to tell the product team that they really seem to have done a good job. However, the UI interface needs to be more intuitive and effective. For a first timer (who could be just a beginner or an experienced data scientist who might be evaluating it for some serious work for his/her organization), the very first experience is not without disappointment even when trying simple things. For example, I wanted to simply try running a decision tree with the IRIS dataset but getting it work really proved to be quite a challenge. I wanted to compare the output of the decision tree predictions of RapidMiner with the model I prepared using Scikit-learn in Python and see how close the two were.

 

Anyway, now that I made it work, I will continue exploring it more and probably some of the early frustrating experience I had with it will probably be compensated with some cool features it has to offer.

 

One thing I would add is the You Tube videos they provided are really of big help. I suggest they should add more such videos to help generate the knowledge about the tool. 

 

 

ggoral

How did you make it work? I got make it work, removing input conection to read csv operator.

 

Thx!

T_Anastasia

You need to import file as data in "repository" menu.
unless it doesn't work.
Why it didn't bound file with your config - I dunno.

T_Anastasia

I mistaked. It still not works for me.
1st file - just master import then press "play"
2nd file - import data then press "play"

varunm1

Hello @T_Anastasia

I see that you imported a file into repository based in the second file you attached. Why do you want to use read CSV as the file is already in your repository? You should drag and drop the file from repository to process window directly, no need to read CSV.

The problem in your earlier process is you connected read CSV to "inp" remove that and you will see an option in read CSV parameters, there you can browse and point to the CSV file in your pc.

T_Anastasia

got it. Thanks