shell

dgraph -dumpsg dumpsg -port 8912

Actors query

We increase the number of actors until we hit just a little below the max.

shell

rm -Rf dumpsg

NUMS="10 56 300"

for NUM in $NUMS; do
  curl localhost:8912/query -XPOST -d "{
          me(_xid_:m.08624h) {
           type.object.name.en
           film.actor.film(first: $NUM) {
               film.performance.film {
                   type.object.name.en
               }
           }
      }
  }" 2>/dev/null | python -m json.tool | wc -l
done

n=0
for S in dumpsg/*.gob; do
  echo $S
  cp -f $S /tmp/actor.${n}.gob
  n=$(($n+1))
done

Directors query

We increase the number of directors until we hit just a little below the max.

shell

rm -Rf dumpsg

NUMS="10 31 100"

for NUM in $NUMS; do
  curl localhost:8912/query -XPOST -d "{
        me(_xid_:m.05dxl_) {
          type.object.name.en
          film.director.film(first: $NUM)  {
                film.film.genre {
                  type.object.name.en
                }
          }
    }
  }" 2>/dev/null | python -m json.tool | wc -l
done

n=0
for S in dumpsg/*.gob; do
  echo $S
  cp -f $S /tmp/director.${n}.gob
  n=$(($n+1))
done

Copy to benchmarks directory

shell

cp -vf /tmp/*.gob ./

Just run run.sh to regenerate these gob files. These gob files are just serialized SubGraph objects.

Just two things to note before running run.sh:

Run it in this directory so that it generates the files here.
Remember to specify DATADIR. This is the location of your posting list data after running dgraph-live-loader.