> -----Original Message-----
> From: Jim Starkey [mailto:jstarkey@stripped]
> Sent: Wednesday, February 18, 2009 3:50 PM
> To: Vladislav Vaintroub
> Cc: Kevin.Lewis@stripped; 'Lars-Erik Bjørk'; 'FalconDev'
> Subject: Re: Patch for bug#42208
>
> It breaks all existing index code. The entire structure of Falcon
> indexes is that the index navigation code treats all keys as naturally
> collated byte arrays.
> Your scheme doesn't have that property, so it would be a total redesign
> and rewrite. Aside from that, your scheme significantly increases the
> code of key comparison. Btrees, you know, are comparison intensive, so
> your scheme would be substantially slower.
>
Jim, you're sure that this does not work. can you please construct an
example where this scheme does not work, specifically Where it breaks
"natural collation"? 2 keys that would sort in RUN scheme differently than
in this scheme? I would not have mentioned it, if the entire structure of
Adabas/Tamino indexes where it was first implemented was so that the index
navigation code treats all keys as naturally collated byte arrays. Exactly
as Falcons.
0 (0x1 0x0) < 1 (0x1 0x01) < any other byte when compared naturally in this
encoding.
> Vladislav Vaintroub wrote:
> > Anyone wants to comment on that?
> > The change would be cheap and fixes at least current multisegment
> padding
> > problem, where 0x00==0x0000==0x000000...
> >
> >
> >> -----Original Message-----
> >> From: Vladislav.Vaintroub@stripped
> [mailto:Vladislav.Vaintroub@stripped]
> >> On Behalf Of Vladislav Vaintroub
> >> Sent: Wednesday, February 18, 2009 12:45 AM
> >> To: Kevin.Lewis@stripped; 'Lars-Erik Bjørk'
> >> Cc: 'Jim Starkey'; 'FalconDev'
> >> Subject: RE: Patch for bug#42208
> >>
> >> Guys, why we're still on alpha stage, maybe we could fix
> multisegments
> >> too?
> >>
> >> Many (i.e 7 or 8) years ago we used following schema for
> multisegment
> >> keys
> >> without padding
> >>
> >> Suppose we have keys A and B and want to make a multisegment key out
> of
> >> it.
> >>
> >> The resulting key would be
> >> f(A) 0x00 f(B)
> >>
> >> 0x00 serves as separator and f() is a transformation that converts
> >>
> >> 0x00=>0x01 0x00
> >> 0x01=>0x01 0x01
> >>
> >> Any other byte remains unchanged.0x00s at the end can be compressed,
> so
> >> we
> >> get efficient key is there are only/many NULLs.
> >>
> >> I do not think the schema is much more complicated than RUN length
> and
> >> padding.
> >>
> >> Vlad
> >>
> >> And while we're on it why not to fix integer representation;)
> Doubles
> >> are
> >> strange creation by mathematician, exact longlongs would be really
> >> nice,
> >> not?
> >>
> >>
> >>
> >>
> >>
> >>> -----Original Message-----
> >>> From: Kevin.Lewis@stripped [mailto:Kevin.Lewis@stripped]
> >>> Sent: Tuesday, February 17, 2009 11:55 PM
> >>> To: Lars-Erik Bjørk
> >>> Cc: Jim Starkey; Vladislav Vaintroub; 'FalconDev'
> >>> Subject: Re: Patch for bug#42208
> >>>
> >>> According to the blog link forwearded by Mark, Oracle customers don
> >>>
> >> not
> >>
> >>> like that zero length strings (which ar equal to each other) are
> >>> automatically converted to NULLs. Both suggestions take care of
> that
> >>> in
> >>> Falcon. So this is the most inportant thing; make a zero length
> >>>
> >> string
> >>
> >>> equal to 0x00 length 1.
> >>>
> >>> The question is whether to keep adding 0x00 to other lengths of
> >>>
> >> binary
> >>
> >>> zero strings. Jim says it does not matter to anyone but QA that
> 0x00
> >>> and 0x0000 sort separately. And Vlad points out that even if we
> did
> >>> this for single field keys, it would not sort them differently for
> >>> multisegment key since we always pad them to a RUN length. I think
> >>> that
> >>> if it does not cause any extra difficulties or comlexity in the
> code,
> >>> why not keep QA happy for single segment keys.
> >>>
> >>> And I still am unclear why this little change in index order should
> >>> cause us to change the ODS format while still in the alpha stage.
> >>>
> >> What
> >>
> >>> is the downside of a new engine that starts converting zero length
> >>> strings into 0x00? New entries will be added to the index after
> the
> >>> NULLS. Older zero length strings would be mixed up with the NULLS
> >>>
> >> and
> >>
> >>> may not be found for direct searched until the index is rebuilt.
> We
> >>> can
> >>> document that as a bug fix in the index, which it is. Nobodies
> >>>
> >> critical
> >>
> >>> data is depending on us finding all zero length strings.
> >>>
> >>> Kevin
> >>>
> >>> Lars-Erik Bjørk wrote:
> >>>
> >>>> Ok, so we probably don't want to do the caching after all then?
> >>>>
> >> Does
> >>
> >>>> anyone else have an opinion on how to proceed on this? Do we agree
> >>>>
> >> on
> >>
> >>>> any best approach?
> >>>>
> >>>> /Lars-Erik
> >>>>
> >>>> Jim Starkey wrote:
> >>>>
> >>>>> Vladislav Vaintroub wrote:
> >>>>>
> >>>>>> Hi Lars-Erik,
> >>>>>> I wonder if adding 0x00 to the (binary) string values that
> >>>>>>
> >> already
> >>
> >>> start
> >>>
> >>>>>> with 0x00 would not be less works that modifying index
> walker
> >>>>>>
> >> etc.
> >>
> >>> This
> >>>
> >>>>>> looks like huge amount of work you have done (good) but I
> wonder
> >>>>>>
> >> if
> >>
> >>>>>> there is
> >>>>>> a good reason for it. Assuming (binary) strings that start
> with
> >>>>>>
> >>> 0x00 are
> >>>
> >>>>>> really seldom, prepending 0x00 to a key after a check is
> not
> >>>>>>
> >> going
> >>
> >>> to
> >>>
> >>>>>> be an
> >>>>>> expensive operation. And that makes NULL *really* different
> from
> >>>>>> other index
> >>>>>> values. And that allows maybe in some distant future
> index-only
> >>>>>> access, so
> >>>>>> you can answer "is null/is not null" without extra accessing
> the
> >>>>>> record and
> >>>>>> this is a real performance advantage.
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>> Why do you want to do that? Is the following sufficient:
> >>>>>
> >>>>> 1. A null is represented as either a zero length key or a
> >>>>>
> >> missing
> >>
> >>>>> segment in a multi-segment key. This collates lowest.
> >>>>> 2. A zero length binary key is represented by a single byte
> of
> >>>>>
> >>> zero.
> >>>
> >>>>> 3. A binary key with a single zero byte is indistinquishable
> >>>>>
> >> from
> >>
> >>> a
> >>>
> >>>>> zero length (but non-null) key
> >>>>> 4. A binary key with a leading zero byte and a subsequent
> non-
> >>>>>
> >> zero
> >>
> >>>>> byte will collate about #2 and #3.
> >>>>>
> >>>>> I don't think we really care about the ordering of a non-null,
> >>>>>
> >> zero
> >>
> >>>>> length key and and all zero binary key. I don't think anyone
> else
> >>>>> should, either.
> >>>>>
> >>>>>
> >>>>
> >> --
> >> Falcon Storage Engine Mailing List
> >> For list archives: http://lists.mysql.com/falcon
> >> To unsubscribe: http://lists.mysql.com/falcon?unsub=1
> >>
> >
> >
> >
>
>
> --
> Falcon Storage Engine Mailing List
> For list archives: http://lists.mysql.com/falcon
> To unsubscribe: http://lists.mysql.com/falcon?unsub=1
