[#4207] - Can't get columns name from CSV datasource correctly

Bug report
Feedback Requested
Project: Severity:
Component: Reproducibility:
Assigned to:

It is no problem in 3.5.2,
however it can't be get correctly in 3.5.3 and 3.6.0.

Is the encoding setting changed from UTF-8 to OS's encoding ?

PS: the csv file's encoding is UTF-8 (Japanese).
the OS is Windows XP SP3 (Japanese)

File C0202002P.csv138 bytes
Image icon 3.5.2_1.GIF21.33 KB
Image icon 3.5.2_2.GIF21.49 KB
Image icon 3.6.0_1-1.GIF21.19 KB
Image icon 3.6.0_1-2.GIF21.23 KB
Image icon 3.6.0_2.GIF21.23 KB
fishlyi's picture
Joined: Aug 3 2009 - 6:11pm
Last seen: 7 years 4 months ago



can you post the file that gives you problems?


The csv file is uploaded.
And it has been confirmed that it has the same problem in 3.6.0.


Unicode characters are not valid as field names, but you can use the special column names for CSV files: COLUMN_1, COLUMN_2, COLUMN_3 and so on... you can add them manually when configuring the connection.
You still have the option to skip the first line by the way.


I verified it works with 3.6.0


feel you free to reopen the bug if you have further questions. Thanks for submitting by the way.


I guess 3.5.3/3.6.0 used not UTF-8 but OS'encoging when it reads the CSV file.
So it maybe works under Linux but not under Windows because the Linux's encoding is UTF-8.


fishlyi, your comment is unclear to me.
I also tested your attached file with iReport 3.6.0 on Windows XP. It works fine both retrieving the column names and running the report and getting the data.

The file that you attached has Windows carriage returns. Your screenshot shows that you select Unix carriage returns as the row delimiter. Perhaps this is the problem.


I'm sorry for my poor English =_=b ...

In 3.5.2, I must select Unix carriage returns(3.5.2_1.gif) so I can get name correctly(3.5.2_2.gif).

In 3.6.0, the result(3.6.0_2.GIF) is the same When Windows carriage returns(3.6.0_1-1.GIF) is selectd, or Unix carriage returns(3.6.0_1-2.GIF) is selected.


I think that I got the reason is the config file [ireport.conf].

From 3.5.3, the config "-J-Dfile.encoding=UTF-8" disappeared,
so it can't read correctly the file of UTF-8 encoding.


Dear list,

i tried using CSV or Excel sources under 3.5 under Win/Lin and now 3.6.1 under Lin/Win. The test goes well, the column detection as well but in the wizard "emptiness".

Does someone has a working example to test if i am stupid?

I hope not, so i reopened the issues.