BaseX = quality
Am Mittwoch, 7. April 2021, 12:25:37 MESZ hat Christian Grün
<christian.gr...@gmail.com> Folgendes geschrieben:
Hi Julia,
Thanks for the comprehensive background information and the references
to our documentation.
Your assumption that the MAXLEN value should be preserved in an
existing database is entirely correct (and the same applies to
MAXCATS). We have fixed that; a new snapshot is available [1,2].
Have fun,
Christian
[1] https://github.com/BaseXdb/basex/issues/2002
[2] https://files.basex.org/releases/latest/
On Tue, Apr 6, 2021 at 4:01 PM Beck, Julia <j.b...@ub.uni-frankfurt.de> wrote:
>
> Dear all,
>
>
> The option MAXLEN tricked me today which is why I need some advice.
>
> I have the problem that I have a query that normally works smoothly but with
> one certain data set, it's really slow. I think it's because at some point in
> the query I am trying to match a value from the index that is cut due to
> MAXLEN of 96 (I made a test: the query is fast for values below 96, but 97
> and above is slow). So I followed the documentation and created the database
> again with a MAXLEN of 200 instead. What tricked me was that whenever I
> optimize the database, MAXLEN is back to 96.
>
>
> The documentation says about MAXLEN:
>
> "Specifies the maximum length for strings to be stored in index structures.
> The value of this option will be assigned once to a new database, and can
> *only be changed* by creating a new database or doing a full optimization."
>
> I assumed "full optimization" means db:optimize with $all is true(). But I
> noticed that the value of MAXLEN also changes when I do db:optimize without
> true().
> For Optimization, the documentation states:
>
> If ALL is specified, all database structures are completely reconstructed.
> The database size will be reduced, and all orphaned data will be deleted.
> Without ALL, only the outdated index structures and database statistics will
> be updated. If the database is completely up-to-date, nothing will be done.
> Database options will be adopted from the original database. Only
> AUTOOPTIMIZE and (if ALL is specified) UPDINDEX will be adopted from the
> current options.
>
> If I understand correctly from the last bullet point, it says that the MAXLEN
> option is not considered from the original database as it's an index option.
> For the second bullet point I would not say though, that "nothing will be
> done", if the database is completely up-to-date. Because when I create a new
> database "test" with a MAXLEN of 200 and do a db:optimize("test"), MAXLEN
> goes back to 96.
> How to keep MAXLEN at the value I set? Do I have to set it every time, I do
> an optimization? Like db:optimize("test", map {"maxlen": 200})? That seems to
> work but it feels a bit over the top, considering that I have many queries
> with optimizations that I would need to adapt for this one data set then.
> What is the reason for resetting MAXLEN during optimization?
>
> Best wishes,
>
> Julia
>
