Option clause very slow

Samur_Araujo · June 8, 2017, 7:47am

Hi all, I observed the option clauses in the query below makes it very slow.

Any suggestion on how to improve it ?

SELECT ?addition ?character ?number ?number_end ?postal_code ?thoroughfare ?bag_id ?city ?city_id ?country ?country_id ?municipality ?municipality_id ?state ?state_id ?lat ?long

WHERE {
 
            VALUES ?address
      {
          <http://geophy.io/addresses/1129>
         
}
 
     ?address address:number ?number .
          ?address address:thoroughfare ?thoroughfare  .
 ?address wgs84:lat ?lat .
          ?address wgs84:long ?long .
 ?address address:city ?city .
          ?city ontologies:id ?city_id .
   ?city gn:locatedIn   ?municipality .
 ?municipality a geotmp:municipality .
         ?municipality ontologies:id ?municipality_id .
   ?municipality gn:locatedIn ?state .
 ?state a geotmp:state .        
 ?state ontologies:id ?state_id .
   ?state gn:locatedIn ?country .
 ?country a geospatial:country .
  
            ?country ontologies:id ?country_id .
OPTIONAL {?address address:addition ?addition .}
OPTIONAL { ?address address:character ?character .}
 OPTIONAL {           ?address address:number_end ?number_end .}
  OPTIONAL {        ?address address:postal_code ?postal_code .}
    OPTIONAL {                ?address tmp:bag_id ?bag_id .}
       

                 }

pavel · June 8, 2017, 9:01am

Hi, it’d be good to know the following:

version of Stardog
query plan with and without the optionals
how much time it takes with and without the optionals.

Do you mean that adding all optionals make it slow or some optionals?

Thanks,
Pavel

Samur_Araujo · June 8, 2017, 11:45am

Query plan with OPTIONALS

$STARDOG/bin/stardog query explain coredb "SELECT ?addition ?character ?number ?number_end ?postal_code ?thoroughfare ?bag_id ?city ?city_id ?country ?country_id ?municipality ?municipality_id ?state ?state_id ?lat ?long WHERE { VALUES ?address { http://geophy.io/addresses/1129 } ?address address:number ?number . ?address address:thoroughfare ?thoroughfare . ?address wgs84:lat ?lat . ?address wgs84:long ?long . ?address address:city ?city . ?city ontologies:id ?city_id . ?city gn:locatedIn ?municipality . ?municipality a geotmp:municipality . ?municipality ontologies:id ?municipality_id . ?municipality gn:locatedIn ?state . ?state a geotmp:state . ?state ontologies:id ?state_id . ?state gn:locatedIn ?country . ?country a geospatial:country . ?country ontologies:id ?country_id . OPTIONAL {?address address:addition ?addition .} OPTIONAL { ?address address:character ?character .} OPTIONAL { ?address address:number_end ?number_end .} OPTIONAL { ?address address:postal_code ?postal_code .} OPTIONAL { ?address tmp:bag_id ?bag_id .} }"
Explaining Query:

SELECT ?addition ?character ?number ?number_end ?postal_code ?thoroughfare ?bag_id ?city ?city_id ?country ?country_id ?municipality ?municipality_id ?state ?state_id ?lat ?long WHERE { VALUES ?address { http://geophy.io/addresses/1129 } ?address address:number ?number . ?address address:thoroughfare ?thoroughfare . ?address wgs84:lat ?lat . ?address wgs84:long ?long . ?address address:city ?city . ?city ontologies:id ?city_id . ?city gn:locatedIn ?municipality . ?municipality a geotmp:municipality . ?municipality ontologies:id ?municipality_id . ?municipality gn:locatedIn ?state . ?state a geotmp:state . ?state ontologies:id ?state_id . ?state gn:locatedIn ?country . ?country a geospatial:country . ?country ontologies:id ?country_id . OPTIONAL {?address address:addition ?addition .} OPTIONAL { ?address address:character ?character .} OPTIONAL { ?address address:number_end ?number_end .} OPTIONAL { ?address address:postal_code ?postal_code .} OPTIONAL { ?address tmp:bag_id ?bag_id .} }

The Query Plan:

From all

Projection(?addition, ?character, ?number, ?number_end, ?postal_code, ?thoroughfare, ?bag_id, ?city, ?city_id, ?country, ?country_id, ?municipality, ?municipality_id, ?state, ?state_id, ?lat, ?long) [#9223372036854.8M]
`─ MergeJoinOuter(?address) [#9223372036854.8M]
   +─ MergeJoinOuter(?address) [#9223372036854.8M]
   │  +─ MergeJoinOuter(?address) [#9223372036854.8M]
   │  │  +─ MergeJoinOuter(?address) [#9223372036854.8M]
   │  │  │  +─ MergeJoinOuter(?address) [#9223372036854.8M]
   │  │  │  │  +─ Sort(?address) [#40679.6M]
   │  │  │  │  │  `─ Bind(<http://geophy.io/addresses/1129> AS ?address) [#40679.6M]
   │  │  │  │  │     `─ LoopJoin(_) [#40679.6M]
   │  │  │  │  │        +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://www.w3.org/2003/01/geo/wgs84_pos#long>, ?long) [#1]
   │  │  │  │  │        `─ LoopJoin(_) [#40679.6M]
   │  │  │  │  │           +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://geophy.io/ontologies/address#number>, ?number) [#1]
   │  │  │  │  │           `─ LoopJoin(_) [#40679.6M]
   │  │  │  │  │              +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://geophy.io/ontologies/address#thoroughfare>, ?thoroughfare) [#1]
   │  │  │  │  │              `─ LoopJoin(_) [#40679.6M]
   │  │  │  │  │                 +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://www.w3.org/2003/01/geo/wgs84_pos#lat>, ?lat) [#1]
   │  │  │  │  │                 `─ HashJoin(?state) [#40679.6M]
   │  │  │  │  │                    +─ MergeJoin(?country) [#2.0M]
   │  │  │  │  │                    │  +─ Sort(?country) [#972]
   │  │  │  │  │                    │  │  `─ NaryJoin(?state) [#972]
   │  │  │  │  │                    │  │     +─ Scan[POSC](?state, rdf:type, <http://geophy.io/ontologies/geotmp#state>) [#972]
   │  │  │  │  │                    │  │     +─ Scan[PSOC](?state, <http://geophy.io/ontologies#id>, ?state_id) [#11.1M]
   │  │  │  │  │                    │  │     `─ Scan[PSOC](?state, <http://www.geonames.org/ontology#locatedIn>, ?country) [#27.7M]
   │  │  │  │  │                    │  `─ MergeJoin(?country) [#252]
   │  │  │  │  │                    │     +─ Scan[POSC](?country, rdf:type, <http://geophy.io/ontologies/geospatial#country>) [#252]
   │  │  │  │  │                    │     `─ Scan[PSOC](?country, <http://geophy.io/ontologies#id>, ?country_id) [#11.1M]
   │  │  │  │  │                    `─ MergeJoin(?municipality) [#415]
   │  │  │  │  │                       +─ Sort(?municipality) [#1]
   │  │  │  │  │                       │  `─ MergeJoin(?city) [#1]
   │  │  │  │  │                       │     +─ Scan[PSOC](?city, <http://www.geonames.org/ontology#locatedIn>, ?municipality) [#27.7M]
   │  │  │  │  │                       │     `─ MergeJoin(?city) [#1]
   │  │  │  │  │                       │        +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://geophy.io/ontologies/address#city>, ?city) [#1]
   │  │  │  │  │                       │        `─ Scan[PSOC](?city, <http://geophy.io/ontologies#id>, ?city_id) [#11.1M]
   │  │  │  │  │                       `─ NaryJoin(?municipality) [#19K]
   │  │  │  │  │                          +─ Scan[POSC](?municipality, rdf:type, <http://geophy.io/ontologies/geotmp#municipality>) [#19K]
   │  │  │  │  │                          +─ Scan[PSOC](?municipality, <http://geophy.io/ontologies#id>, ?municipality_id) [#11.1M]
   │  │  │  │  │                          `─ Scan[PSOC](?municipality, <http://www.geonames.org/ontology#locatedIn>, ?state) [#27.7M]
   │  │  │  │  `─ Scan[PSOC](?address, <http://geophy.io/ontologies/address#addition>, ?addition) [#532K]
   │  │  │  `─ Scan[PSOC](?address, <http://geophy.io/ontologies/address#character>, ?character) [#898K]
   │  │  `─ Scan[PSOC](?address, <http://geophy.io/ontologies/address#number_end>, ?number_end) [#74K]
   │  `─ Scan[PSOC](?address, <http://geophy.io/ontologies/address#postal_code>, ?postal_code) [#5.4M]
   `─ Scan[PSOC](?address, <http://geophy.io/ontologies/tmp#bag_id>, ?bag_id) [#5.4M]

Samur_Araujo · June 8, 2017, 11:46am

Query plan without OPTIONALS:

samur@samur-XPS-15-9550:~/development/coredb-ontologies$ $STARDOG/bin/stardog query explain coredb "SELECT ?addition ?character ?number ?number_end ?postal_code ?thoroughfare ?bag_id ?city ?city_id ?country ?country_id ?municipality ?municipality_id ?state ?state_id ?lat ?long WHERE { VALUES ?address { http://geophy.io/addresses/1129 } ?address address:number ?number . ?address address:thoroughfare ?thoroughfare . ?address wgs84:lat ?lat . ?address wgs84:long ?long . ?address address:city ?city . ?city ontologies:id ?city_id . ?city gn:locatedIn ?municipality . ?municipality a geotmp:municipality . ?municipality ontologies:id ?municipality_id . ?municipality gn:locatedIn ?state . ?state a geotmp:state . ?state ontologies:id ?state_id . ?state gn:locatedIn ?country . ?country a geospatial:country . ?country ontologies:id ?country_id . }"
Explaining Query:

SELECT ?addition ?character ?number ?number_end ?postal_code ?thoroughfare ?bag_id ?city ?city_id ?country ?country_id ?municipality ?municipality_id ?state ?state_id ?lat ?long WHERE { VALUES ?address { http://geophy.io/addresses/1129 } ?address address:number ?number . ?address address:thoroughfare ?thoroughfare . ?address wgs84:lat ?lat . ?address wgs84:long ?long . ?address address:city ?city . ?city ontologies:id ?city_id . ?city gn:locatedIn ?municipality . ?municipality a geotmp:municipality . ?municipality ontologies:id ?municipality_id . ?municipality gn:locatedIn ?state . ?state a geotmp:state . ?state ontologies:id ?state_id . ?state gn:locatedIn ?country . ?country a geospatial:country . ?country ontologies:id ?country_id . }

The Query Plan:

From all

Projection(?addition, ?character, ?number, ?number_end, ?postal_code, ?thoroughfare, ?bag_id, ?city, ?city_id, ?country, ?country_id, ?municipality, ?municipality_id, ?state, ?state_id, ?lat, ?long) [#68.1M]
`─ Bind(<http://geophy.io/addresses/1129> AS ?address) [#68.1M]
   `─ LoopJoin(_) [#68.1M]
      +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://www.w3.org/2003/01/geo/wgs84_pos#long>, ?long) [#1]
      `─ LoopJoin(_) [#68.1M]
         +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://geophy.io/ontologies/address#number>, ?number) [#1]
         `─ LoopJoin(_) [#68.1M]
            +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://geophy.io/ontologies/address#thoroughfare>, ?thoroughfare) [#1]
            `─ LoopJoin(_) [#68.1M]
               +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://www.w3.org/2003/01/geo/wgs84_pos#lat>, ?lat) [#1]
               `─ HashJoin(?state) [#68.1M]
                  +─ MergeJoin(?country) [#2.0M]
                  │  +─ Sort(?country) [#972]
                  │  │  `─ NaryJoin(?state) [#972]
                  │  │     +─ Scan[POSC](?state, rdf:type, <http://geophy.io/ontologies/geotmp#state>) [#972]
                  │  │     +─ Scan[PSOC](?state, <http://geophy.io/ontologies#id>, ?state_id) [#11.1M]
                  │  │     `─ Scan[PSOC](?state, <http://www.geonames.org/ontology#locatedIn>, ?country) [#27.7M]
                  │  `─ MergeJoin(?country) [#252]
                  │     +─ Scan[POSC](?country, rdf:type, <http://geophy.io/ontologies/geospatial#country>) [#252]
                  │     `─ Scan[PSOC](?country, <http://geophy.io/ontologies#id>, ?country_id) [#11.1M]
                  `─ MergeJoin(?municipality) [#1]
                     +─ Sort(?municipality) [#1]
                     │  `─ MergeJoin(?city) [#1]
                     │     +─ Scan[PSOC](?city, <http://www.geonames.org/ontology#locatedIn>, ?municipality) [#27.7M]
                     │     `─ MergeJoin(?city) [#1]
                     │        +─ Scan[SPOC](<http://geophy.io/addresses/1129>, <http://geophy.io/ontologies/address#city>, ?city) [#1]
                     │        `─ Scan[PSOC](?city, <http://geophy.io/ontologies#id>, ?city_id) [#11.1M]
                     `─ NaryJoin(?municipality) [#19K]
                        +─ Scan[POSC](?municipality, rdf:type, <http://geophy.io/ontologies/geotmp#municipality>) [#19K]
                        +─ Scan[PSOC](?municipality, <http://geophy.io/ontologies#id>, ?municipality_id) [#11.1M]
                        `─ Scan[PSOC](?municipality, <http://www.geonames.org/ontology#locatedIn>, ?state) [#27.7M]

zachary.whitley · June 8, 2017, 1:45pm

You might be interested in these three blog posts

pavel · June 9, 2017, 8:20am

Hi Samur,

Thanks for the plans. I have one guess what might cause the slowness but it’s hard to tell without data or further information. Could you answer the remaining questions and also tell how many results the query without the optionals return?

Also, we’d really appreciate if you send plans in an attached text file next time, that’d preserve the formatting.

Cheers,
Pavel

system · June 23, 2017, 8:20am

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.

Topic		Replies	Views
Query still slow after optimize db Support	8	386	February 4, 2021
Stardog rule is making a 50 ms query take 500 ms Bug	6	645	March 29, 2017
Stardog performance issue in Stardog 7 Bug	2	376	August 22, 2019
Query Performance Has Changed Support	5	339	October 21, 2019
Query performance issues Support	8	1158	February 28, 2017

Option clause very slow

Related Topics