Question

I have a huge xml file whose sample data is as follows :

 <vendor name="aglaia"><br>
              <vendorOUI oui="000B91" description="Aglaia Gesellschaft für Bildverarbeitung ud Kommunikation m" /><br>
         </vendor><br>
         <vendor name="ag"><br>
              <vendorOUI oui="0024A9" description="Ag Leader Technology" /><br>
         </vendor><br>

as it can be see there are text ” Gesellschaft für Bildverarbeitung ” which is not UTF-8 compliant because which I am getting errors from the xml validator , errors like:

Import failed:
com.sun.org.apache.xerces.internal.impl.io.MalformedByteSequenceException: Invalid byte 1 of 1-byte UTF-8 sequence.

So the query is how to take care of this in Linux environment to convert the xml file to UTF-8 compliant format? or is there a way in bash such that while creating the xml in the first place i can ensure that all variables/strings are stored in UTF-8 compliant format?

Answer 1

Use the character set conversion tool:

iconv -f ISO-8859-1 -t UTF-8 filename.txt

See gnu-page

…and in file http://standards.ieee.org/develop/regauth/oui/oui.txt “aglia” (as in your example above) is reported as:

00-0B-91   (hex)            Aglaia Gesellschaft für Bildverarbeitung und Kommunikation m
000B91     (base 16)        Aglaia Gesellschaft für Bildverarbeitung und Kommunikation m
                            Tiniusstr. 12-15
                            Berlin  D-13089
                            GERMANY

it seems like “ü” is the character that gets mangeld.

Update

When downloading “oui.txt” using wget, I see the character “ü” in the file. If you don’t have that something is broken in your download. consider using one of these:

• wget --header='Accept-Charset: utf-8'
• try using curl -o oui.txt instead

If none of the above works, just open the link in you favorite browser and do a “save as”. In that case, comment the wget line in the script below.

I had success with the following script (update BEGIN & END to get a valid XML-file)

#!/bin/bash

wget http://standards.ieee.org/develop/regauth/oui/oui.txt
iconv -f iso-8859-15 -t utf-8 oui.txt > converted

awk 'BEGIN {
         print "HTML-header"
     }

     /base 16/ {
         printf("<vendor name=\"%s\">\n", $4)
         read
         desc = substr($0, index($0, $4))
         printf("<vendorOUI oui=\"%s\" description=\"%s\"/>\n", $1, desc)
     }
     END {
         print "HTML-footer"
    }
    ' converted

Hope this helps!