Re: [bdbxml] Varying XQuery performance with change in 'return' clause

> There are a couple of ways to tackle this - one is to try a query like th=
is:
>
> for $x in collection("genbank_bct.dbxml")/INSDSet/INSDSeq
> let $y :=3D INSDSeq_feature-table/INSDFeature/INSDFeature_quals/
>    INSDQualifier/INSDQualifier_value
> where $y =3D "3.2.1.1"
> return $x/INSDSeq_locus
>

So, I executed two queries. One was

for $x in collection("genbank_bct.dbxml")/INSDSet/INSDSeq
let $y :=3D $x/INSDSeq_feature-table/INSDFeature/INSDFeature_quals/INSDQual=
ifier/INSDQualifier_value
where $y=3D"3.2.1.1"
return
$x/INSDSeq_locus

which was exactly the same as yours. This query took a long time to
execute. I stopped counting after about 10 mins. Its query plan was as
follows :
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
<XQuery>
  <FLWOR>
    <ForBinding name=3D"x">
      <Navigation>
        <QueryPlanFunction result=3D"collection" container=3D"genbank_bct.d=
bxml">
          <OQPlan>U</OQPlan>
        </QueryPlanFunction>
        <Step axis=3D"child" name=3D"INSDSet" nodeType=3D"element"/>
        <Step axis=3D"child" name=3D"INSDSeq" nodeType=3D"element"/>
      </Navigation>
    </ForBinding>
    <LetBinding name=3D"y">
      <Navigation>
        <Variable name=3D"x"/>
        <Step axis=3D"child" name=3D"INSDSeq_feature-table" nodeType=3D"ele=
ment"/>
        <Step axis=3D"child" name=3D"INSDFeature" nodeType=3D"element"/>
        <Step axis=3D"child" name=3D"INSDFeature_quals" nodeType=3D"element=
"/>
        <Step axis=3D"child" name=3D"INSDQualifier" nodeType=3D"element"/>
        <Step axis=3D"child" name=3D"INSDQualifier_value" nodeType=3D"eleme=
nt">
          <OQPlan>P(edge-element-equality-string,prefix,INSDQualifier.INSDQ=
ualifier_value)</OQPlan>
        </Step>
      </Navigation>
      <Where>
        <Navigation>
          <Variable name=3D"y"/>
          <DbXmlCompare name=3D"equal">
            <OQPlan>V(edge-element-equality-string,INSDQualifier.INSDQualif=
ier_value,=3D,'3.2.1.1')</OQPlan>
            <Sequence>
              <AnyAtomicTypeConstructor value=3D"3.2.1.1"
typeuri=3D"http://www.w3.org/2001/XMLSchema" typename=3D"string"/>
            </Sequence>
          </DbXmlCompare>
        </Navigation>
      </Where>
    </LetBinding>
    <Navigation>
      <Variable name=3D"x"/>
      <Step axis=3D"child" name=3D"INSDSeq_locus" nodeType=3D"element">
        <OQPlan>P(unique-edge-element-equality-string,prefix,INSDSeq.INSDSe=
q_locus)</OQPlan>
      </Step>
    </Navigation>
  </FLWOR>
</XQuery>
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D

Then, I executed the following query :
for $x in collection("genbank_bct.dbxml")/INSDSet/INSDSeq/INSDSeq_feature-t=
able/INSDFeature/INSDFeature_quals/INSDQualifier[INSDQualifier_value=3D"3.2=
.1.1"]
return
$x/../../../../INSDSeq_locus

This query came back in about 4 mins as I reported earlier. The query
plan was as follows :
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
<XQuery>
  <Navigation>
    <QueryPlanFunction result=3D"collection" container=3D"genbank_bct.dbxml=
">
      <OQPlan>n(P(unique-edge-element-equality-string,prefix,INSDSeq.INSDSe=
q_locus),V(edge-element-equality-string,INSDQualifier.INSDQualifier_value,=
=3D,'3.2.1.1'))</OQPlan>
    </QueryPlanFunction>
    <Step axis=3D"child" name=3D"INSDSet" nodeType=3D"element"/>
    <Step axis=3D"child" name=3D"INSDSeq" nodeType=3D"element"/>
    <Step axis=3D"child" name=3D"INSDSeq_feature-table" nodeType=3D"element=
"/>
    <Step axis=3D"child" name=3D"INSDFeature" nodeType=3D"element"/>
    <Step axis=3D"child" name=3D"INSDFeature_quals" nodeType=3D"element"/>
    <Step axis=3D"child" name=3D"INSDQualifier" nodeType=3D"element"/>
    <DbXmlFilter>
      <Navigation>
        <Step axis=3D"child" name=3D"INSDQualifier_value" nodeType=3D"eleme=
nt"/>
        <DbXmlCompare name=3D"equal">
          <Sequence>
            <AnyAtomicTypeConstructor value=3D"3.2.1.1"
typeuri=3D"http://www.w3.org/2001/XMLSchema" typename=3D"string"/>
          </Sequence>
        </DbXmlCompare>
      </Navigation>
    </DbXmlFilter>
    <Step axis=3D"parent" uri=3D"*" name=3D"*" nodeType=3D"*"/>
    <Step axis=3D"parent" uri=3D"*" name=3D"*" nodeType=3D"*"/>
    <Step axis=3D"parent" uri=3D"*" name=3D"*" nodeType=3D"*"/>
    <Step axis=3D"parent" uri=3D"*" name=3D"*" nodeType=3D"*"/>
    <Step axis=3D"child" name=3D"INSDSeq_locus" nodeType=3D"element"/>
  </Navigation>
</XQuery>
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D

So, I guess it doesnt help as far as reducing the query times is
concerned. I would like to know what exactly is happening in the first
case.

Can I try something else to bring the query times below 4 mins that is
the current best.

Thanks
Deepak


------------------------------------------
To remove yourself from this list, send an
email to [email protected]