postgres: How to skip the fist and the last row of a queryset in a for-loop?

Question

In a function I'm looping over the all rows a select statement returns. I want to skip the fist and the last row. How could I do this?

...
  FOR arow IN
    SELECT (dp).path[1] As index, 
           ST_AsText((dp).geom) As wktnode, 
           (dp).geom As vertex
      FROM (SELECT 
          ST_DumpPoints(
              ST_Segmentize(NEW.the_geom,
                  -- the max length of the segment 
                  ST_Length(
                      NEW.the_geom
                  )/NEW.dist_calc_pnts
              )
          ) AS dp
    ) As foo
    LOOP

      test_me := arow.index;
      IF NOT (ST_Equals(arow.vertex, ST_StartPoint(NEW.the_geom))
                      OR
             ST_Equals(arow.vertex, ST_EndPoint(NEW.the_geom))) THEN

      RAISE NOTICE 'Calling update_pntzzz(%)', test_me;
      INSERT INTO 
         pntzzz(onedee_id, the_geom)
      VALUES(NEW.id, arow.vertex);
      END IF;
    END LOOP;
...

As I said, I do not need the result from first and the last iteration but I'm struggling on how to implement this. Right now the if statement checks whether the geometry of the current row is identical to the start or endpoint of the original line geometry - a rather expensive check.

Answer 1

This is some sample code to show how to skip the first and the last element in a loop.

It should be simple to adapt it to your use case.

DO $$DECLARE
   counter integer := 0;
   i integer;
   old_i integer;
BEGIN
   FOR i IN 1..10 LOOP
      counter := counter + 1;
      IF counter > 2 THEN
         -- processing of old_i goes here
         RAISE WARNING 'Processing row %', old_i;
      END IF;

      old_i := i;
   END LOOP;
END;$$;