just a suggestion

Alex, I have to admit I like this script more and more :)

for the XML import: it might be a nice idea to get the columns by grabbing the name-tags from the first record in the xml file.

For the following XML-code the Column-names would be automatically generated to:
column1 = filename; column2 = actie and so on.

<?xml version="1.0" ?>
- <root>
- <record>
<filename>bout.prt</filename>
<actie>ChangeValue</actie>
<parameter>LENGTH</parameter>
<value>45</value>
<opmerking>een verklaring</opmerking>
</record>
- <record>
<filename>moer.prt</filename>
<actie>ChangeValue</actie>
<parameter>DIA</parameter>
<value>23</value>
<opmerking>een verklaring</opmerking>
</record>

Yes, I will try to implement something like this. However it is far from obvious when you have more complex XML structure. Should I list attributes as well or elements only? What if I have nested elements? What if 'records' have different structure?

Also I believe in real life the display labels and XML tags are nearly always different, especially if you start thinking of internationalization etc.

Alex

yes, you are right, it is a hard job to create a standard for this, if a standard could be made in the first place.

maybe it is a nice command for the simple xml only, to read the element-names for the header.
to exploit this for the less simple xml-formats is not that easy :)