add genpinyin2.awk

diff --git a/scripts/Makefile.data b/scripts/Makefile.data
new file mode 100644 (file)
index 0000000..2f665b2
--- /dev/null
+++ b/scripts/Makefile.data
@@ -0,0 +1,9 @@
+all: pinyin2.txt
+
+
+pinyin2.txt:
+       awk -f genpinyin2.awk ../data/gb_char.table > $@
+
+
+.PHONY: pinyin2.txt
+

diff --git a/scripts/genpinyin2.awk b/scripts/genpinyin2.awk
new file mode 100644 (file)
index 0000000..00c6cad
--- /dev/null
+++ b/scripts/genpinyin2.awk
@@ -0,0 +1,8 @@
+#!/usr/bin/awk
+       { if (length($2) == 2) pinyins[$1] += $4 }
+
+END {
+    for (pinyin in pinyins) {
+        print pinyin, pinyins[pinyin]
+    }
+}