lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From dar...@ontrenet.com
Subject RE: Python Response Bug?
Date Thu, 25 Jun 2009 16:38:54 GMT
The problem is what should Solr put for a float that is undefined? There
is no such value. Typically a value anomaly should not exist and so when
data structures break, they reveal these situations. The correct action is
to explore why a non-value is making its way into the index and correct it
before that by applying your own logic to populate the field with a
tangible value.

> Regardless, I think it should return valid JSON so programs don't crash
> when trying to interpret it. I don't think about these things often so
> maybe I'm missing something obvious, but I think putting in an empty
> string is better than putting in nothing and having it break.
>
> My 2 cents.
> Mike
>
>
> -----Original Message-----
> From: darren@ontrenet.com [mailto:darren@ontrenet.com]
> Sent: Thursday, June 25, 2009 11:11 AM
> To: solr-user@lucene.apache.org
> Cc: solr-user@lucene.apache.org
> Subject: Re: Python Response Bug?
>
> The first JSON is invalid as you see because of the missing value.
>
> :, is not valid JSON syntax.
>
> The reason it works for string type is because the empty string '' is a
> valid field value but _nothing_ (the first examepl) is not. There is no
> empty float placeholder that JSON likes. So either it has to be an empty
> string or some legitimate float value.
>
> FAIK, Solr doesn't generate empty numeric values compatible with JSON.
> zero and null are different things.
>
>
>
>> I'm not sure, but I think I ran across some unexpected behavior in the
>> python response in solr 1.3 (1.3.0 694707 - grantingersoll -
> 2008-09-12
>> 11:06:47).
>>
>> I'm running Python 2.5 and using eval to convert the string solr
> returns
>> to python data objects.
>>
>> I have a blank field in my xml file that I am importing labeled
>> "contacthours". If I make the field a "string" type, the interpreter
>> doesn't throw an error. If I make the field a "float" type, python
>> throws an error on the eval function. Here is portion of the output
> from
>> solr, with the two differences near the end by the "contacthours"
>> variable:
>>
>>
>> Broken (Python didn't like this):
>>
> {'responseHeader':{'status':0,'QTime':0,'params':{'q':'id:"161"','wt':'p
>>
> ython'}},'response':{'numFound':1,'start':0,'docs':[{'approvalpending':'
>> ','assessmentcourseprograminstitutionimprovement':u'The students\u2019
>> final reports along with the \u201cBiology Externship Completion
>> Form\u201d, which is completed by externship supervisors, will be used
>> for this assessment.','contacthourrationale':'Variable hours - no
>> explanation given','contacthours':,'coursearea':'BIO',
>>
>>
>> Good (this worked):
>>
> {'responseHeader':{'status':0,'QTime':0,'params':{'q':'id:"161"','wt':'p
>>
> ython'}},'response':{'numFound':1,'start':0,'docs':[{'approvalpending':'
>> ','assessmentcourseprograminstitutionimprovement':u'The students\u2019
>> final reports along with the \u201cBiology Externship Completion
>> Form\u201d, which is completed by externship supervisors, will be used
>> for this assessment.','contacthourrationale':'Variable hours - no
>> explanation given','contacthours':'','coursearea':'BIO',
>>
>>
>> Any insights? Is this a bug or am I missing something?
>>
>> Mike Beccaria
>> Systems Librarian
>> Head of Digital Initiatives
>> Paul Smith's College
>> 518.327.6376
>> mbeccaria@paulsmiths.edu
>>
>>
>>
>
>


Mime
View raw message