Implement the -a and -v options in join.awk.

[dxcommon.git] / scripts / join.awk

diff --git a/scripts/join.awk b/scripts/join.awk
index 548ca2cd3fe955c2674d34b767d6e071cb6a87c7..7187b50c3e1d048c9d1bc2066088af6323694e62 100755 (executable)
--- a/scripts/join.awk
+++ b/scripts/join.awk
@@ -2,43 +2,115 @@
 #
 # Copyright © 2022 Nick Bowler
 #
-# Partial implementation of POSIX "join" command.  No options are supported.
+# Partial implementation of POSIX "join" command: only the "-v" and "-a"
+# options are implemented.
+#
+# Not all awk implementations support reading from standard input with the
+# getline function by specifying a filename of "-".  In particular, busybox
+# awk will read from a file named "-" instead of standard input.  Since
+# busybox-based environments are typically missing "join", this limitation
+# is problematic.  As a workaround, do not use "-" for the second input
+# filename.
 #
 # License WTFPL2: Do What The Fuck You Want To Public License, version 2.
 # This is free software: you are free to do what the fuck you want to.
 # There is NO WARRANTY, to the extent permitted by law.
 
 BEGIN {
-  if (ARGC != 3) {
+  show_uniq_lhs = 0
+  shoq_uniq_rhs = 0
+  show_common   = 1
+
+  # Process command-line options
+  for (i = 1; i < ARGC; i++) {
+    if (substr(ARGV[i], 1, 1) != "-" || ARGV[i] == "-")
+      break;
+
+    opt = substr(ARGV[i], 2, 1);
+    if (opt == "a" || opt == "v") {
+      num = substr(ARGV[i], 3, 1);
+      if (num == "") {
+        # option argument must be next on command-line
+        ARGV[i++] = "";
+        num = ARGV[i];
+      }
+
+      if (opt == "v") {
+        show_common = 0;
+      }
+
+      if (num == 1) {
+        show_uniq_lhs = 1;
+      } else if (num == 2) {
+        show_uniq_rhs = 1;
+      } else {
+        # invalid argument
+        exit 1;
+      }
+    } else {
+      # unsupported option
+      exit 1;
+    }
+
+    ARGV[i] = "";
+  }
+
+  if (i+2 != ARGC) {
+    # invalid usage
     exit 1;
   }
 
-  file2 = ARGV[2];
-  delete ARGV[2];
+  file2 = ARGV[i+1];
+  ARGV[i+1] = "";
 
-  advance_rhs();
+  rhs_max_nr = rhs_nr = 0;
+  if (advance_rhs() == 0)
+    finish_rhs();
 }
 
+{ $1 = $1 }
 $1 == lhs_prev {
-  # Rewind RHS as we have duplicate keys in LHS.
+  # Rewind RHS as we have duplicate common keys in LHS.
   close(file2);
+
+  rhs_nr = 0;
   advance_rhs();
 }
 
-$1 < rhs[1] { next }
+$1 < rhs[1] {
+  if (show_uniq_lhs) {
+    print;
+  }
+  next;
+}
+
 {
   while ($1 > rhs[1]) {
+    if (show_uniq_rhs && rhs_nr == rhs_max_nr)
+      print_rhs();
+
     if (advance_rhs() == 0)
-      exit(0);
+      finish_rhs();
+
+    if (show_uniq_lhs && $1 < rhs[1])
+      print;
   }
 }
 
-$1 == rhs[1] {
-  lhs_prev = $1 = $1;
+!rhs_eof && $1 == rhs[1] {
+  lhs_prev = $1;
   do {
-    print_match();
+    if (show_common) print_match();
     advance_rhs();
-  } while ($1 == rhs[1]);
+  } while (!rhs_eof && $1 == rhs[1]);
+}
+
+END {
+  if (show_uniq_rhs) {
+    do {
+      print_rhs();
+    } while (advance_rhs() > 0);
+  }
 }
 
 function advance_rhs(raw, rc)
@@ -47,15 +119,43 @@ function advance_rhs(raw, rc)
   if (rc < 0)
     exit(1);
 
+  rhs_eof = rc == 0;
+
+  if (rhs_max_nr == rhs_nr++)
+    rhs_max_nr++;
+
   split(raw, rhs);
   return rc;
 }
 
+function finish_rhs(rc)
+{
+  rc = 0;
+  if (show_uniq_lhs) {
+    do {
+      if (NR > 0) {
+        $1 = $1;
+        print;
+      }
+    } while ((rc = getline) > 0);
+  }
+  exit(-rc);
+}
+
+function print_rhs(i)
+{
+  if (!rhs_eof) {
+    if (i < 1)
+      printf "%s", rhs[1];
+    for (i = 2; i in rhs; i++) {
+      printf " %s", rhs[i];
+    }
+    print "";
+  }
+}
+
 function print_match(i)
 {
   printf "%s", $0
-  for (i = 2; i in rhs; i++) {
-    printf " %s", rhs[i];
-  }
-  print ""
+  print_rhs(2);
 }