Separate v2 and v3 into dedicated directories

[khatus.git] / v2 / src / awk / lib / num.awk

diff --git a/v2/src/awk/lib/num.awk b/v2/src/awk/lib/num.awk
new file mode 100644 (file)
index 0000000..4c22f9d
--- /dev/null
+++ b/v2/src/awk/lib/num.awk
@@ -0,0 +1,37 @@
+function num_round(n) {
+    return int(n + 0.5)
+}
+
+function num_ensure_numeric(n) {
+    return n + 0
+}
+
+#-----------------------------------
+# Why do we need num_ensure_numeric?
+#-----------------------------------
+# awk appears to be guessing the type of an inputted scalar based on usage, so
+# if we read-in a number, but did not use it in any numeric operations, but did
+# use as a string (even in just a format string!) - it will be treated as a
+# string and can lead to REALLY SURPRISING behavior in conditional statements,
+# where smaller number may compare as greater than the bigger ones, such as.
+#
+# Demo:
+#
+# $ 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"}}'
+# 75 < 100
+# $ 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"}}'
+# 75 > 100
+
+# However, once used as a number, seems to stay that way even after being
+# used as string:
+#
+# $ 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"}}'
+# 75 < 100
+# 
+# $ 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"}}'
+# 75 < 100
+# 
+# $ awk 'BEGIN {x = "75"; y = "100"; x + y; z = x y; if (x > y) {print "75 > 100"} else if (x < y) {print "75 < 100"}}'
+# 75 < 100
+# $ awk 'BEGIN {x = "75"; y = "100"; z = x y; if (x > y) {print "75 > 100"} else if (x < y) {print "75 < 100"}}'
+# 75 > 100