#common-tips

1)this doesnt work

(q (as emp :e1)
   (join (q (as emp :e2)
            ((not= :e1.ename :e2.ename)
             (>= :e2.hiredate :e1.hiredate))
            (group {:next-hiredate (min :e2.hiredate)}))
         (true? true))
   show)
   
e1 is not visable in the nested q
its not like mongodb i cant see the e1, solution to this is do the join with e1,e2 in the creteria ONLY
and after the join is done to d the subquery

2)sort is allowed in array but not on Map types

(group :e1.ename :e1.hiredate {:next-hiredate (min [:e2.hiredate :e2.ename])})

BUT ARRAY REQUIRES SAME TYPES SO PROBLEM

3)conj-each accumulator, doesnt collect nulls 


4)(q (as emp :e1)
   [{:clerks (if- (= :job "CLERK") :ename nil)}]
   {:rn (wfield (row-number) (wsort :clerks!))}
   
  e1.rn doesnt exists, its added later
  
5)window field + row number

  row_number and then group
  or rn+group are useful alot
  
  to make groups that didnt exist before
  
6)if i want to group and avoid to collect on array (maybe data too many)
  i can group by more fields to make each raw unique (i can add row number also)
  
  
7)if i want to put 2 tables one next to another with join
  i make them to be same row-count, i use row-number
  and i join based on row-number,
  because row-number cant be always the same number, i do  fullouter join