firstinfield and position attributes usage

[Henri-Damien]

Hi,
I have problems using firstinfield and position attributes and then
querying with @attr 3=1.
Indeed, some users would find it cool to search anatomy only with the
beginning of the phrase.


First of all, I wonder how I can define a custom index and use it.
It seems that using completeness attribute implies @attr 6=1 if 0 @attr
6=2 or @attr 6=1.
position is not defined in you documentation.
firstinfield seems to be used along with it.

I tried to define :

index h
completeness 0
position 1
firstinfield 1
charmap sort-string-utf.chr

in default.idx and
used it in record.abs :
melm 250$a Heading, Heading:h, Heading:s

and queried my base as such :
f @and @attr "1=Authority/format-id" @attr 5=100 MSC  @attr "1=Heading"
@attr 3=1 An

Answer was :
ERROR 119

Is there something I missed ?

--
Henri Damien LAURENT et Paul POULAIN
Consultants indépendants
en logiciels libres et bibliothéconomie (http://www.koha-fr.org)



_______________________________________________
Koha-zebra mailing list
Koha-zebra@...
http://lists.nongnu.org/mailman/listinfo/koha-zebra

[Sebastian Hammer]

Henri-Damien LAURENT wrote:

>Hi,
>I have problems using firstinfield and position attributes and then
>querying with @attr 3=1.
>Indeed, some users would find it cool to search anatomy only with the
>beginning of the phrase.
>
>
>First of all, I wonder how I can define a custom index and use it.
>It seems that using completeness attribute implies @attr 6=1 if 0 @attr
>6=2 or @attr 6=1.
>position is not defined in you documentation.
>firstinfield seems to be used along with it.
>  
>

Hi Henri,

Zebra at present does not support the first-in-field attribute, although
we sometimes have talked about doing so. You can simulate it pretty well
by using a combination of complete subfield and truncation. You'll note
there's also a way to configure an index to ignore leading articles
(the, an, a, etc.), which is a good way to increase your recall.. yes,
it's hard to make this word with word boundaries, (Adam might have a
suggestion for some regexp magic that would work), but I think that's
probably a minor problem in most cases.

--Sebastian

>I tried to define :
>
>index h
>completeness 0
>position 1
>firstinfield 1
>charmap sort-string-utf.chr
>
>in default.idx and
>used it in record.abs :
>melm 250$a Heading, Heading:h, Heading:s
>
>and queried my base as such :
>f @and @attr "1=Authority/format-id" @attr 5=100 MSC  @attr "1=Heading"
>@attr 3=1 An
>
>Answer was :
>ERROR 119
>
>Is there something I missed ?
>
>  
>

--
Sebastian Hammer, Index Data
quinn@...   www.indexdata.com
Ph: [phone] Fax: [phone]



_______________________________________________
Koha-zebra mailing list
Koha-zebra@...
http://lists.nongnu.org/mailman/listinfo/koha-zebra

[Adam Dickmeiss]

Henri-Damien LAURENT wrote:

> Hi,
> I have problems using firstinfield and position attributes and then
> querying with @attr 3=1.
> Indeed, some users would find it cool to search anatomy only with the
> beginning of the phrase.
>
>
> First of all, I wonder how I can define a custom index and use it.
> It seems that using completeness attribute implies @attr 6=1 if 0 @attr
> 6=2 or @attr 6=1.
> position is not defined in you documentation.
> firstinfield seems to be used along with it.
>
> I tried to define :
>
> index h
> completeness 0
> position 1
> firstinfield 1
> charmap sort-string-utf.chr
>
> in default.idx and
> used it in record.abs :
> melm 250$a Heading, Heading:h, Heading:s
>
> and queried my base as such :
> f @and @attr "1=Authority/format-id" @attr 5=100 MSC  @attr "1=Heading"
> @attr 3=1 An
>
> Answer was :
> ERROR 119

Your index named 'h' is not used for "normal" searches. The structure
and sometimes the completeness attribute tell _what_ index type to use.
If no structure is given .. and completeness is omitted or set to 1
(incomplete subfield) then index 'w' is used. Not your index 'h'. The
error message suggests that firstinfield is not enabled for your index 'w'.

You can either just enable firstinfield for 'w' .. It's already enabled
  for 'w' in tab/default.idx that comes with Zebra 2.0. But probably you
don't use that.

You can also just specify that you want to use index 'h' in your query
by using structure=h  (@attr 4=h). This structure hack requires Zebra
2.0.14.

/ Adam

>
> Is there something I missed ?
>



_______________________________________________
Koha-zebra mailing list
Koha-zebra@...
http://lists.nongnu.org/mailman/listinfo/koha-zebra