PT-1572 WIP

bin/pt-table-checksum

@@ -13262,6 +13262,10 @@ the values are converted to strings by the CONCAT() function, and MySQL chooses
 the string representation.  If you specify a value of 2, for example, then the
 values 1.008 and 1.009 will be rounded to 1.01, and will checksum as equal.
 
+=item --force-concat-enums
+
+bla bla bla bla bla
+
 =item --function
 
 type: string

lib/NibbleIterator.pm

@@ -134,14 +134,37 @@ sub new {
       );
       PTDEBUG && _d('Ascend params:', Dumper($asc));
 
+      # Check if enum fields items are sorted or not.
+      # If they are sorted we can skip adding CONCAT to improve the queries eficiency.
+      my $force_concat_enums = $o->has('force-concat-enums') && $o->get('force-concat-enums');
+      for my $index (@{$index_cols}) {
+          if ($tbl->{tbl_struct}->{type_for}->{$index} eq 'enum') {
+             if ($tbl->{tbl_struct}->{defs}->{$index} =~ m/enum\s*\((.*?)\)/) {
+                 my @items = split(/,\s*/, $1);
+                 my $sorted = 1; # Asume the items list is sorted to later check if this is true
+                 for (my $i=1; $i < scalar(@items); $i++) {
+                     if ($items[$i-1] gt $items[$i]) {
+                         $sorted = 0;
+                         last;
+                     }
+                 }
+                 if (!$force_concat_enums && !$sorted) {
+                    die "The index " . $index . " in table " .  $tbl->{name} .
+                    " has unsorted enum items.\nPlease read the documentation for the --force-concat-enums parameter\n";
+                 }
+             }
+          }
+      }
+
       # Make SQL statements, prepared on first call to next().  FROM and
       # ORDER BY are the same for all statements.  FORCE IDNEX and ORDER BY
       # are needed to ensure deterministic nibbling.
+
       my $from     = "$tbl->{name} FORCE INDEX(`$index`)";
-      my $order_by = join(', ', map { $tbl->{tbl_struct}->{type_for}->{$_} eq 'enum' 
+      my $order_by = join(', ', map { $tbl->{tbl_struct}->{type_for}->{$_} eq 'enum' && $force_concat_enums
                                         ? "CONCAT(".$q->quote($_).")" : $q->quote($_)} @{$index_cols});
 
-      my $order_by_dec = join(' DESC,', map { $tbl->{tbl_struct}->{type_for}->{$_} eq 'enum' 
+      my $order_by_dec = join(' DESC,', map { $tbl->{tbl_struct}->{type_for}->{$_} eq 'enum' && $force_concat_enums
                                         ? "CONCAT(".$q->quote($_).")" : $q->quote($_)} @{$index_cols});
 
       # The real first row in the table.  Usually we start nibbling from

t/lib/NibbleIterator.t

@@ -38,6 +38,8 @@ my $dbh = $sb->get_dbh_for('master');
 
 if ( !$dbh ) {
    plan skip_all => 'Cannot connect to sandbox master';
+} else {
+    plan tests => 60;
 }
 
 my $q   = new Quoter();
@@ -879,7 +881,7 @@ is_deeply(
 # #############################################################################
 
 diag(`/tmp/12345/use < $trunk/t/lib/samples/cardinality.sql >/dev/null`);
-
+$dbh->do('analyze table bad_tables.inv');
 $ni = make_nibble_iter(
    db   => 'cardb',
    tbl  => 't',
@@ -892,6 +894,56 @@ is(
    "Use non-unique index with highest cardinality (bug 1199591)"
 );
 
+$sb->load_file('master', "t/lib/samples/NibbleIterator/enum_keys.sql");
+$ni = undef;
+eval {
+    $ni = make_nibble_iter(
+       db   => 'test',
+       tbl  => 't1',
+       argv => [qw(--databases test --chunk-size 3)],
+    );
+};
+
+like(
+    $EVAL_ERROR,
+    qr/The index f3 in table `test`.`t1` has unsorted enum items/,
+    "PT-1572 Die on unsorted enum items in index",
+);
+
+eval {
+    $ni = make_nibble_iter(
+       db   => 'test',
+       tbl  => 't1',
+       argv => [qw(--databases test --force-concat-enums --chunk-size 3)],
+    );
+};
+
+like( 
+    $ni->{explain_first_lb_sql},
+    qr/ORDER BY `f1`, `f2`, CONCAT\(`f3`\)/,
+    "PT-1572 Use of CONCAT for unsorted ENUM field items without --",
+);
+
+eval {
+    $ni = make_nibble_iter(
+       db   => 'test',
+       tbl  => 't2',
+       argv => [qw(--databases test --chunk-size 3)],
+    );
+};
+
+is(
+    $EVAL_ERROR,
+    '',
+    "PT-1572 No errors on sorted enum items in index",
+);
+
+like( 
+    $ni->{explain_first_lb_sql},
+    qr/ORDER BY `f1`, `f2`, `f3`/,
+    "PT-1572 Don't use CONCAT for sorted ENUM field items without --force-concat-enums",
+);
+
 # #############################################################################
 # Done.
 # #############################################################################
@@ -905,6 +957,7 @@ like(
    qr/Complete test coverage/,
    '_d() works'
 );
+
 $sb->wipe_clean($dbh);
 ok($sb->ok(), "Sandbox servers") or BAIL_OUT(__FILE__ . " broke the sandbox");
 done_testing;

t/lib/samples/NibbleIterator/enum_keys.sql

@@ -0,0 +1,47 @@
+DROP DATABASE IF EXISTS test;
+CREATE DATABASE test;
+USE test;
+
+-- Don't change the comments. The enum word inside the comment is there to test the table parser
+
+CREATE TABLE `test`.`t1` ( 
+    f1 DATE NOT NULL,
+    f2 INT(10) UNSIGNED NOT NULL,
+    f3 ENUM('c','a','b','d') NOT NULL DEFAULT 'c' COMMENT "unsorted enum items",
+    f4 INT(10) UNSIGNED NOT NULL DEFAULT '0',
+    PRIMARY KEY (`f1`,`f2`,`f3`) 
+) ENGINE=InnoDB DEFAULT CHARSET=utf8;
+
+
+INSERT INTO `t1` VALUES 
+('2018-07-27',165910161,'c',1375471152),
+('2018-07-27',393713658,'d',1382302491),
+('2018-07-27',472875023,'c',525456967),
+('2018-07-27',543582931,'c',1657080267),
+('2018-07-27',583532949,'d',280366509),
+('2018-07-27',1396416465,'d',1252007743),
+('2018-07-27',1705409249,'c',1714682759),
+('2018-07-27',1801160058,'a',1022430181),
+('2018-07-27',1898674299,'c',1310715836),
+('2018-07-27',2011751560,'a',109015753);
+
+CREATE TABLE `test`.`t2` ( 
+    f1 DATE NOT NULL,
+    f2 INT(10) UNSIGNED NOT NULL,
+    f3 ENUM('a','b','c','d') NOT NULL DEFAULT 'c' COMMENT "sorted enum items",
+    f4 INT(10) UNSIGNED NOT NULL DEFAULT '0',
+    PRIMARY KEY (`f1`,`f2`,`f3`) 
+) ENGINE=InnoDB DEFAULT CHARSET=utf8;
+
+
+INSERT INTO `t2` VALUES 
+('2018-07-27',165910161,'c',1375471152),
+('2018-07-27',393713658,'d',1382302491),
+('2018-07-27',472875023,'c',525456967),
+('2018-07-27',543582931,'c',1657080267),
+('2018-07-27',583532949,'d',280366509),
+('2018-07-27',1396416465,'d',1252007743),
+('2018-07-27',1705409249,'c',1714682759),
+('2018-07-27',1801160058,'a',1022430181),
+('2018-07-27',1898674299,'c',1310715836),
+('2018-07-27',2011751560,'a',109015753);