Number of pairs in calculating Jaccard distance using PySpark are less than they should be

Question

I am trying to calculate Jaccard distance between certain ids with their attributes in the form of SparseVectors.

from pyspark.ml.feature import MinHashLSH
fr