CSV Writer prepending question mark "?" to column name

I am unable to figure out why the CSV writer is outputing one column name with a question mark prepended. This only is happening with 1 of 4 columns, and the data values for that column are unaffected. The column output looks like "?ColumnName". Other writers, like an ExcelWriter, do not cause this issue and output the column names correctly. When I change the encoding from the default on the CSVWriter to UTF-8, the "?" changes to a "". Where is this strange character coming from? I have simplified my workspace such that I have a File Reader followed by the CSV Writer, and still the column name is output incorrectly. The source file has no spurious characters and is ASCII encoded. Viewing the FileTable of the File Reader shows the column with the correct name as well. Other CSV Writers I had in the same workspace on other sets of data output just fine.

There may be a related issue. The CSV Writer's output file with the bad column name seems to have corrupt data in one of the other columns preventing Oracle Data Visualization from loading that other column as a date. The data loads fine as a String, but I cannot convert it to a date. This suggests that this other column may be getting a non-printing character concatenated with the values as well, but I haven't dug into the data yet to try and find it. Again, outputing the same data using an Excel Writer doesn't corrupt the country name, and the date column values are interpreted correctly by Oracle Data Visualization.

Thoughts? Thanks!

Hi Frutchey,

could you post me a screenshot?

Thanks! Iris

I found a workaround to this issue. Seems that the default quote characters are being added to every non-numerical value, and the quote character is not being recognized by Excel or Oracle Data Visualization periodically (usually the values are interpreted correctly, other times they get corrupted). By removing the quotes from the output everything is working.

This is still strange, as other data output with the same quotes around non-numeric values work just fine. Attaching screenshots and my source data file if someone wants to keep debugging...


FileTable for reader:

Output file: