教算法用awk比PASCAL强

请看KMP代码：

function KMP(s, m, p, n) {
  if (n == 0 || m == 0) return;
  split(s, str, ""); split(p, pattern, "");
  # https://www.tutorialspoint.com/awk/awk_string_functions.htm
  # split返回的数组，[0]是""，所以下面是PASCAL版
  # https://blog.csdn.net/idler/article/details/4846
  j = 1; k = 0;
  while (j < n) {
    if ( k == 0 || pattern[j] == pattern[k] ) {
        j = j + 1; k = k + 1;
        na[j] = k; /* next是关键字 */
    }
    else k = na[k];
  }
  for (i = j = 1; i <= m;) {
    # 不支持,表达式，不支持//
    if ( j == 0 || str[i] == pattern[j] ) { i++; ++j; }
    else j = na[j];
    if (j > n) { printf("%d\n", i - n - 1); j = 1; }
  }
}
BEGIN {
  # awk doen't support variables local to functions
  _str = "AABACAABAAAABAA12345AABAA";
       _ptn = "AABAA";
  m = length(_str); n = length(_ptn); # m长得像nn, 所以比较长
  KMP(_str, m, _ptn, n);
}

数组是关联数组/map:

BEGIN {
  str = "a b\tc  de" # 语句不用;结尾
  split(str, ary, "[ \t]+") # C还得用外部库
  for (i in ary) printf("%d %s\n", i, ary[i])
  print # 比puts(""); 短
  for (i = 0; i < length(ary); i++) printf("%d %s\n", i, ary[i])
  ary["tom"] = "cat"; ary[3.14] = 5678 # 关联数组/map
  print; print(ary["tom"], ary[3.14])
}

MadEdit有语法高亮：

gawk -f t.awk

https://files.cnblogs.com/files/blogs/714801/gawk.7z 176KB 含且仅含 gawk.exe 461KB