Some clarification on the matter.
The file in question has xml content, tags and all.
The behaviour described previously also occurs with html. The
following code was used:
<!DOCTYPE html>
<html>
<body>
<table>
<tr>
<td>100</td>
<td>200</td>
<td>300</td>
</tr>
<tr>
<td>400</td>
<td>500</td>
<td>600</td>
</tr>
</table>
</body>
</html>
It used to be imported into a calc sheet if one opens it as a html calc
document.
Was this import directly from the bank server?
Or was the bank data copied from the web browser into calc?
If the html file above is copied from the web browser (opera) into
calc, the dialogue window 'text import' appears.
The calc sheet looked like a normal sheet in an .ods file.
Now it opens in something remotely resembling a text table in writer
although in the document window it says libreoffice.calc .
And acordingly when you try to save it as something else , the save as
dialogue doesn't have an option for .ods format.
Seems logical behaviour. Html is an authoring language, so writer is
the appropriate module for such code.
I would expect if I attempt to open an .html table in calc - the table to
get imported into a spreadsheet,
and if I attempt to open an .html table in writer - the table to be opened
as a text table (document),
which is not the case anymore as far as calc is concerned.
Personally, disagree. Html code should open in writer. Where html
contains tabular data, why should calc be expected to interpret and
convert non-numeric attributes such as spacing, width, etc.?
In writer, you should be able to select the table (e.g. using the
'navigator' dialogue window), then select the tabular data, then copy
to calc?