Xapian discuss - Apr 2007 - Re: Xapian-discuss Digest, Vol 35, Issue 9

On 4/9/07, xapian-discuss-request@lists.xapian.org
<xapian-discuss-request@lists.xapian.org> wrote:
> Message: 6
> Date: Mon, 9 Apr 2007 13:53:15 -0700
> From: "Mark Blythe" <list@markblythe.com>
> Since Xapian doesn't have any expectations about the data you store in
the
> document, you can store whatever you like.  It could be a serialized data
> structure (using Data::Dumper, JSON, etc), or something more simple like a
> delimited string.  In your case, you might store something like:
Oh dear ... for real? Hmmm. Well, if that's the only way ... still,
i'd rather not have to write a parser to get back display results.
Since any character could be allowed in the description, it's an
unsafe assumption to say a delitimited string would work 100% of the
time. And using eval to reconstruct a Data::Dumper dumped hash ref can
be dangerous. XML? Great. :|

Thanks for help ...

-- 
jeffa

I wouldn't go so far as to write a full parser either, but there are plenty
of pre-existing packages for serializing/deserializing data structures.  You
shouldn't have to re-invent the wheel.  If you want to go with a delimited
string and can't think of any printable character that you can guarantee
won't be in the data, you can always choose a non-printable delimiter, like
a chr(0) or chr(7).  If you want to store key/value pairs but don't trust
eval Data::Dumper, use JSON.  It's eval-safe and language-independent.

By not enforcing a key/value data setup, Xapian stays completely flexible
and can support storing any kind of data, including BLOBs, etc.  I
personally think it was a smart choice for a generic library.

On 4/9/07, Jeff Anderson <captvanhalen@gmail.com>
wrote:
>
> On 4/9/07, xapian-discuss-request@lists.xapian.org
> <xapian-discuss-request@lists.xapian.org> wrote:
> > Message: 6
> > Date: Mon, 9 Apr 2007 13:53:15 -0700
> > From: "Mark Blythe" <list@markblythe.com>
> > Since Xapian doesn't have any expectations about the data you
store in
> the
> > document, you can store whatever you like.  It could be a serialized
> data
> > structure (using Data::Dumper, JSON, etc), or something more simple
like
> a
> > delimited string.  In your case, you might store something like:
>
> Oh dear ... for real? Hmmm. Well, if that's the only way ... still,
> i'd rather not have to write a parser to get back display results.
> Since any character could be allowed in the description, it's an
> unsafe assumption to say a delitimited string would work 100% of the
> time. And using eval to reconstruct a Data::Dumper dumped hash ref can
> be dangerous. XML? Great. :|
>
> Thanks for help ...
>
> --
> jeffa
>
> _______________________________________________
> Xapian-discuss mailing list
> Xapian-discuss@lists.xapian.org
> http://lists.xapian.org/mailman/listinfo/xapian-discuss
>