KNIME version could be the problem
KNIME version could be the problem
Hi; maybe it’s a version difference, but it’s a strange situation. I haven’t been able to find the cause yet, I tried different things, but the result is negative. @bruno29a I’ll take a look if you share the version and settings.
My KNIME 4.3.4
Hi @umutcankurt , my default encoding setting is Cp1252. Yours is set to Cp1254, which is different:
That could explain it
EDIT: I set mine to Cp1254, and it still displays properly in my Excel Reader…
Hi; @bruno29a Thank you for the answer. Any idea how I can solve this situation? And your setting (workspace) screen editor picture, also what is the knime version ?
Hi @umutcankurt my settings is like yours, except that it shows Cp1252. But you can add whichever you want via the “Other” option.
For my Knime version, it’s the latest, 4.4.1.
I’m not sure how to solve this for you. Not sure if it’s Knime related, or system related. If it is the latter, upgrading to 4.4.1 will not change resolve this.
@umutcankurt I tried to read the data with the Excel Reader under Windows but it failed. With MacOSX and several nodes I think I can read the data and also it seems there are R packages that can read them as CSV and store the results in other formats like KNIME .table or ARFF or Parquet or SQLite which you might later use.
The (new) CSV reader although had a problem also under MacOSX.
So while R might not be the easiest option there could be a workaround.
Hi @mlauber71 ,
Do you mean it could not read the file, or that it was not displaying the strings correctly?
Which version of Knime are you running? I’m asking because I was able to open the file with the Excel Reader under Windows and it displayed properly too (please check my screenshots).
Why would it work for some and not for others…?
It did not display the correct characters under a Windows 10 system. Not sure why that is. I did experiment with Differenzen UTF settings. But to no avail.
Thanks for the additional info @mlauber71 . I’m running on Windows 8.1.
The Excel Reader does not seem to offer character encoding option.
Are you able to open the file via Excel, copy the content and paste in 1) Knime’s Table Creator; and 2) A text file via notepad? Do the text display properly?
Hi, there is something missing or strange somehow, but I’m sure I couldn’t find it about the version or a little setting. I hope you guys can find the reason for this slight difference.
@bruno29a I can open the file in Excel and copy the characters (so I was able to extract them and also have them in a ‘pure’ table). You can follow that in the /data/ folder of my example.
The Excel Reader does not read the characters neither in Windows 10 nor MacOSX. R does the job under Mac. I was not able to test it in Windows since KNIME gives strange messages about the R version (sad).
My; Windows 10 Home Single Language and KNIME 4.3.4
Which encoding did you use? I was not able to read it on window 10
Hi; @mlauber71 and @bruno29a
Airport.xlsx (52.5 KB)
In order to filter through the word list, I translate and update many languages in google tables. I use it to filter the data by importing the word list into knime. I upgraded the knime version, I tried different ones, but I couldn’t get it to display in 3 languages. Could you please help me solve the problem by examining the attached excel file. What am I missing? I can not find.
My knime version 4.4.1
Sysytem Windows 10
@umutcankurt I tried some variations also with different characters, Under MacOS all seems to work fine. I still saw strange problems under Windows 10. It could display the characters in R/Rstudio (to some extent).
One thing to try next would be to import the data from excel directly via R, but I fear there is a display problem under Windows 10 within some (all?) nodes.
All 41 columns are displaying well.
I can confirm that the characters are displayed wrongly in the table view on Win10 - I created a ticket for that (internal reference: AP-17532), so hopefully it will be resolved in the next release.
@umutcankurt unfortunately I have no idea for a workaround . Maybe its merely a displaying error and the nodes work as they should, despite the wrong display?
@all: Thank you so much for the thorough testing!
I tried several workarounds with R nodes via CSV, Excel and Parquet files. The problem always is that some characters are not displayed correctly within nodes in KNIME.
- bn (Bengali)
- hi (Hindi) although that should have worked in the past
- ko (korean)
All these characters should be part of a ‘regular’ UTF-8 encoding though …
Exactly, they don’t get displayed correctly, but I think they get read in and handled correctly - you can still configure and apply other nodes with these special characters with the expected results:
e.g. filtering or applying a rule engine (the rule I used:
$3$ = "বিমানবন্দর" => "gotcha!", displayed as
). So in principle you can build your workflow, only type the respective characters in an editor that displays them nicely, then copy&paste them to where you need them.
But yeah, even it is “only” a displaying issue: it is ugly and annoying and needs to be fixed Where the root caus is? I have no idea, because as you said and found out, these characters are part of UTF-8 and they get displayed properly in the annotations… maybe the font in the dialogues is lacking these characters and thus displays an emtpy box? but that’s only speculation
Thank you for the temporary solution. But since I need to add many words, it is not possible for me to edit it manually. I hope the new knime version will be fixed. Again, thank you all for the research and all the suggestions.