performance - ConstantScoreQuery in elasticsearch -

i'm not sure how constantscorequery works regarding performance in elasticsearch 1.4 (i don't think changed between 1.0 , 2.0 regarding this)

a simple query has calculate score if query :

{    "constant_score":     "match": {"field":"text"} } 

it's faster without constant_score.

now if doing boolquery, there difference between wrapping boolquery constant_score or wrapping every must/should/must_not clause ? exemples :

{  "constant_score": {   "bool":{     "must": [         {"match": {"field1": "text1"}},         {"match": {"field2": "text2"}}     ]   }  } } 

versus

{   "bool":{     "must": [         {"constant_score": {"match": {"field1": "text1"}}},         {"constant_score": {"match": {"field2": "text2"}}}     ]   } } 

another one, when boolfilter 2 "should" cases, default "minimum_should_match" set 1 , if first tested case ok doesn't bother check other.

with queries, score important each "should" tested , adds score of hit. there way constant_score wrapping bool query make simple "or" bool above filter ? mean not checking every "should" on each document.

thanks insights :)

i tested on local elasticsearch (1.4.4)

it turns out if wrap constant_score around bool query shoulds doesn't behave boolfilter, returns every document. minimum should match mandatory.

about using constant_score on bool query or on each part, return same, every result has same score still don't know if computed before , set constant one...