Forums / Extensions / eZ Find / Potential BC break for subattribute fields in eZ Find 2.3: input requested
Paul Borgermans
Wednesday 16 June 2010 3:16:27 am
While field names in Solr are in principle fully unicode, there are some areas where hyphens (or dash '-') pose problems. This is used in particular with subattribute field names.
I want to change this convention and replace the hyphen with an underscore ('_').
Apart from a re-index need, this will break cases where you use hardcoded fields names in templates for example when using subattribute filtering.
I would like to have an idea how many of you use hard coded subattribute fields and your comments on this change
Regards
Paul
eZ Publish, eZ Find, Solr expert consulting and training http://twitter.com/paulborgermans
Nicolas Pastorino
Wednesday 16 June 2010 3:34:28 am
Hi Paul and thanks for polling the community !
My two cents on the topic : The "-" (hyphen) separator was chosen to make a visual difference from the other separator used by eZ Find when building Solr field names : "_" (underscore). This helps a lot any developer to identify/qualify a field's origin, namely when reading direct outputs from Solr (console, Solr admin).
In short, having two different separators adds, in my opinion, a valuable semantic nuance, and i'd recommend choosing another, Solr-compliant character than "_" for this purpose.
Cheers,
-- Nicolas Pastorino Director Community - eZ Member of the Community Project Board eZ Publish Community on twitter: http://twitter.com/ezcommunity t : http://twitter.com/jeanvoye G+ : http://plus.tl/jeanvoye
Wednesday 16 June 2010 4:12:01 am
The point is to avoid special characters which are actually also part of the lucene query syntax:
+ - && || ! ( ) { } [ ] ^ " ~ * ? : \
An alternatve is to use a pattern with multiple characters to aid human readability
For example
_SUB_
as a separator
More thoughts?
Ivo Lukac
Wednesday 16 June 2010 6:36:58 am
We have one project using this feature, put it will not be a problem to upgrade :)
Cheers.
http://www.linkedin.com/in/ivolukac http://www.netgen.hr/eng/blog http://twitter.com/ilukac
Wednesday 16 June 2010 1:41:19 pm
ok, last chance to influence me:
what about '___' which are three consecutive underscores?
examples:
submeta_related_article___class_name_s subattr_related_article___body_t
Is that clear enough to see the attribute and subattribute identifiers?
gilles guirand
Wednesday 16 June 2010 7:50:07 pm
Enjoy the three consecutive underscores !
-- Gilles Guirand eZ Community Board Member http://twitter.com/gandbox http://www.gandbox.fr
Wednesday 16 June 2010 11:40:30 pm
+1