On 30-12-2008 13:35:13 +0100, Stefan Manegold wrote:
On Tue, Dec 30, 2008 at 12:38:54PM +0100, Wouter Alink wrote:
I haven't looked at your document, but it is probably the byte-order mark...
See section: 3.2 Byte order mark (BOM) http://www.ietf.org/rfc/rfc2781.txt
Normal editors/viewers hide it, as it isn't a visible character. I think MonetDB should have stripped it before showing you the error message.
If "magic numbers" and/or "byte order marks" are indeed to be expected in *plain* text (csv) files, then yes, ortherwise, it could only mention that the file does not comply to the expected *plain* text format.
Stefan's €0.02 ;-)
Could argue that since we only support UTF-8, we should try to support UTF-8 files in all kinds of flavours, or not? % file ~/Desktop/ImportedCSV0.csv ~/Desktop/ImportedCSV0.csv: UTF-8 Unicode (with BOM) text, with CRLF line terminators