#!/afs/athena/contrib/perl/perl
$RCS_ID = '$Id: summ,v 2.2 1992/06/23 08:42:14 hobbs Exp $' ;
($pgm  = $0) =~ s-.*/-- ;
$HelpInfo = <<EOH ;

	    RDB operator: $pgm

Usage:  $pgm  [options]  [column ...]

Options:
    -cu      A Count of the unique values for each column given.
    -cun     Like option '-cu' but also shows counts of null (empty) and
	     blank values (have only space chars), if either exist.
    -cuu     A Count of each unique value for each column given.
    -cu2     Like option '-cuu' but shows only counts greater than one.
    -help    Print this help info.
    -m       The min, average, max, total for each column given.
    -v	     Inverse option. Selects all columns except those named.

Produces "summary" information about the rdbtable. If no columns are given
then information about all columns is produced.  A Count of the data rows
is always shown.

This operator reads an rdbtable via STDIN and writes a summary report via
STDOUT.  Options may be abbreviated.

$RCS_ID
EOH
while ( $ARGV[0] =~ /^-/ ) {				# Get args
    $_ = shift ;
    if( /-c(.*)/ ){
	if( $1 =~ /^u/ ){ $CUNIQ++ ; $SAV++ ; }
	if( $1 =~ /^un/ ){ $CUNUL++ ; }
	if( $1 =~ /^uu/ ){ $CUQBY++ ; }
	if( $1 =~ /^u2/ ){ $CUQBY++ ; $CU2++ ; }
	next ; }
    if( /-h.*/ ){ print $HelpInfo ; exit; }
    if( /-m(.*)/ ){ $MAM++ ; $SAV++ ; next ; }
    if( /^-v.*/ ){ $INV++ ; next ; }
    die "Bad arg: $_\n", $HelpInfo ;
}
while(<STDIN>){
    next if ! $lln && /^#\s|^\s+#/ ;	# comment
    chop ;
    @F = split( /\t/, $_ );
    if( ++$lln < 3 ){
	if( $lln == 1 ){			# col name line
	    @H = @F ; # save headers
	    @ARGV = @H if ! @ARGV && ! $INV ;
	    &get_col_x ; }		# get, chk column indexes.
	next ; }
    $rows++ ;				# data row count
    if( $SAV ){
	for (@a){
	    $x = "$_|$F[$_]" ;
	    if( ! $q{$x} ){ $cu[$_]++ ; }	# unique count
	    $q{$x}++ ;				# value count
	}
    }
}
print "Rows: $rows\n" ;
if( $CUNIQ ){				# count of unique stuff
    for (@a){
	print "Unique values for $H[$_]: $cu[$_]\n" ;
	next if ! $CUQBY && ! $CUNUL ;
	for $k ( sort( keys %q )){
	    @t = split( /\|/, $k );
	    if( $t[0] eq $_ ){
		if( $t[1] eq "" ){ $t[1] = "(null)" ; }
		if( $t[1] =~ /^\s+$/ ){ $t[1] = "(blank)" ; }
		if( $CUNUL ){
		    last if $t[1] ne "(null)" && $t[1] ne "(blank)" ; }
		if( $CU2 ){
		    next unless $q{$k} > 1 ; }
		printf( "%8d %s\n", $q{$k}, $t[1] ) ;
	    }
	}
    }
}
if( $MAM ){				# min, avg, max 
    for (@a){
	$n = $sum = $max = 0 ;
	$min = 2e31 -1 ;
	for $k ( sort( keys %q )){
	    @t = split( /\|/, $k );
	    if( $t[0] eq $_ ){
		$sum += $t[1] * $q{$k} ;
		$n += $q{$k} ;
		$min = $t[1] if $t[1] < $min ;
		$max = $t[1] if $t[1] > $max ;
	    }
	}
	$avg = $sum / $n if $n ;
	printf( "Min, Avg, Max, Total for %s: %d, %d, %d, %d\n",
	    $H[$_], $min, $avg, $max, $sum ) ;
    }
}
sub get_col_x {		# get, chk column indexes, inc -v, die if bad column
			# uses @H, $INV, put indexes in @a.
			# modified for summ.
    local( $f, $ok, @nn ) ;
    for $arg (@ARGV){
	for( $ok=$f=0 ; $f < @H ; $f++ ){
	    if( $arg eq $H[$f] ){	# match existing column
		$ok++ ;
		push( @a, $f );
		last ; }
	}
	die "$0: Bad column name: $arg\n" if ! $ok ;
    }
    if( $INV ){					# inverse option
	loop: for( $f=0 ; $f < @H ; $f++ ){
	    for $i (@a){
		next loop if $i eq $f ; }
	    push( @nn, $f ); }
	@a = @nn ;
    }
}