Re: Spreadsheet translation

[Philip Nienhuis]

"Markus Mützel" wrote:
> Am 31. Januar 2020 um 10:31 Uhr schrieb "Philip Nienhuis":
> > Well, OCT can (I wrote the code for it) but .xlsx is a very very
> > complicated spreadsheet format (basically it's a compressed nested
> > directory tree full of XML files), and we read it using regular
> > expressions while XML had better be read using XML parsers and such.
> > In fact, all the Java based spreadsheet I/O classes are based on XML
> > parsers/validators etc. But regexps are much faster than XML (although I
> > often think OTOH they're much more fragile).
>
> This legendary stackoverflow answer comes to mind ;-)
> https://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags/1732454#1732454
>
> Markus

Thanks Markus, I'll add that link to the wiki.
An amusing read - further down the arguments pro and con from the strict 
vs. loose/practical developer types makes for more interesting reading 
moments :-)

As to our issue, it's especially the order of XML tags in the cell nodes 
that makes our regexps fragile. Luckily, both Excel and LibreOffice 
write those tags in similar order (~alphabetical, AFAICT). But maybe 
some Java based spreadsheet I/O class or embedded SW or so will do 
otherwise, and then we might have a bit of trouble. I'm unsure if there 
are rules for the order of appearance of tags in XML.
OTOH reading and writing just simple data, which Octave does, isn't so 
terribly complicated and demanding.

Philip