numberic or string type for non-sortable field?

2011-03-01 Thread cyang2010
I wonder if i shall use solr int or string for such field with following
requirement

multi-value
facet needed
sort not needed


The field value is a an id.  Therefore, i can store as either numeric field
or just a string.   Shall i choose string for efficiency?

Thanks.

-- 
View this message in context: 
http://lucene.472066.n3.nabble.com/numberic-or-string-type-for-non-sortable-field-tp2606353p2606353.html
Sent from the Solr - User mailing list archive at Nabble.com.


Re: numberic or string type for non-sortable field?

2011-03-01 Thread Ahmet Arslan
 I wonder if i shall use solr int or
 string for such field with following
 requirement
 
 multi-value
 facet needed
 sort not needed
 
 
 The field value is a an id.  Therefore, i can store as
 either numeric field
 or just a string.   Shall i choose string
 for efficiency?

Trie based integer (tint) is preferred for faster faceting.
 





Re: numberic or string type for non-sortable field?

2011-03-01 Thread Chris Hostetter

:  The field value is a an id.  Therefore, i can store as
:  either numeric field
:  or just a string.   Shall i choose string
:  for efficiency?
: 
: Trie based integer (tint) is preferred for faster faceting.

range faceting/filtering yes -- not for field faceting which is what i 
think he's asking about.

in that case int would still proably be more efficient, but you don't want 
precision steps (that will introduce added terms)

-Hoss

Re: numberic or string type for non-sortable field?

2011-03-01 Thread cyang2010
Sorry i didn't make my question clear.

I will only facet based on field value, not ranged query  (it is just some
ids for a  multi-value field).   And i won't do sort on the field either.

In that case, is string more efficient for the requirement?

-- 
View this message in context: 
http://lucene.472066.n3.nabble.com/numberic-or-string-type-for-non-sortable-field-tp2606353p2606762.html
Sent from the Solr - User mailing list archive at Nabble.com.


Re: numberic or string type for non-sortable field?

2011-03-01 Thread Ahmet Arslan
 I will only facet based on field value, not ranged
 query  (it is just some
 ids for a  multi-value field).   And i
 won't do sort on the field either.
 
 In that case, is string more efficient for the
 requirement?

Hoss was saying to use, fieldType name=int class=solr.TrieIntField 
precisionStep=0 omitNorms=true positionIncrementGap=0/ 





Re: numberic or string type for non-sortable field?

2011-03-01 Thread cyang2010
Can I know why?  I thought solr is tuned for string if no sorting of facet by
range query is needed.

-- 
View this message in context: 
http://lucene.472066.n3.nabble.com/numberic-or-string-type-for-non-sortable-field-tp2606353p2607932.html
Sent from the Solr - User mailing list archive at Nabble.com.


Re: numberic or string type for non-sortable field?

2011-03-01 Thread Chris Hostetter

: Can I know why?  I thought solr is tuned for string if no sorting of facet by
: range query is needed.

tuned for string doesn't really mean anything to me, i'm not sure what 
that's in refrence to.  nothing thta i know of is particularly optimized 
for strings.  Almost anything can be indexed/stored/represented as a 
string (in some form ot another) and that tends to work fine in solr, but 
some things are optimized for other more specialized datatypes.

the reason i suggested that using ints might (marginally) be better is 
because of the FieldCache and the fieldValueCache -- the int 
representation uses less memory then if it was holding strings 
representing hte same ints.

worrying about that is really a premature optimization though -- model 
your data in the way that makes the most sense -- if your ids are 
inherently ints, model them as ints until you come up with a reason to 
model them otherwise and move on to the next problem.


-Hoss