This test used indexRandom to index 4 documents. indexRandom introduces also bogus documents which were getting on the way now that we use fieldValueFactor instead of a script to determine the script of the documents. Taken out indexRandom to simplify things and make the tests more predictable.