awk/lib/num.awk

   1 function num_round(n) {
   2     return int(n + 0.5)
   3 }
   4
   5 function num_ensure_numeric(n) {
   6     return n + 0
   7 }
   8
   9 #-----------------------------------
  10 # Why do we need num_ensure_numeric?
  11 #-----------------------------------
  12 # awk appears to be guessing the type of an inputted scalar based on usage, so
  13 # if we read-in a number, but did not use it in any numeric operations, but did
  14 # use as a string (even in just a format string!) - it will be treated as a
  15 # string and can lead to REALLY SURPRISING behavior in conditional statements,
  16 # where smaller number may compare as greater than the bigger ones, such as.
  17 #
  18 # Demo:
  19 #
  20 # $ awk 'BEGIN {x = "75"; y = "100"; sprintf("x: %d, y: %d\n", x, y); if (x > y) {print "75 > 100"} else if (x < y) {print "75 < 100"}}'
  21 # 75 < 100
  22 # $ awk 'BEGIN {x = "75"; y = "100"; sprintf("x: %s, y: %d\n", x, y); if (x > y) {print "75 > 100"} else if (x < y) {print "75 < 100"}}'
  23 # 75 > 100
  24
  25 # However, once used as a number, seems to stay that way even after being
  26 # used as string:
  27 #
  28 # $ awk 'BEGIN {x = "75"; y = "100"; x + y; sprintf("x: %s, y: %d\n", x, y); if (x > y) {print "75 > 100"} else if (x < y) {print "75 < 100"}}'
  29 # 75 < 100
  30 #
  31 # $ awk 'BEGIN {x = "75"; y = "100"; x + y; sprintf("x: %s, y: %d\n", x, y); z = x y; if (x > y) {print "75 > 100"} else if (x < y) {print "75 < 100"}}'
  32 # 75 < 100
  33 #
  34 # $ awk 'BEGIN {x = "75"; y = "100"; x + y; z = x y; if (x > y) {print "75 > 100"} else if (x < y) {print "75 < 100"}}'
  35 # 75 < 100
  36 # $ awk 'BEGIN {x = "75"; y = "100"; z = x y; if (x > y) {print "75 > 100"} else if (x < y) {print "75 < 100"}}'
  37 # 75 > 100