| 1 |
use strict; |
| 2 |
use Symbol; |
| 3 |
|
| 4 |
my $vardir = $ENV{MYSQLTEST_VARDIR} |
| 5 |
or die "need MYSQLTEST_VARDIR"; |
| 6 |
|
| 7 |
# fixed parameters |
| 8 |
|
| 9 |
my $par_nullpct = 10; # pct nulls if nullable |
| 10 |
my $par_sppct = 5; # a few pct special values like min,max |
| 11 |
my $par_quotepct = 20; # pct quoted if fields_enclosed_by |
| 12 |
my $par_valerrpct = 2 ; # pct value errors |
| 13 |
my $par_randbias = 3; # prefer smaller random values |
| 14 |
|
| 15 |
# type info |
| 16 |
|
| 17 |
sub gen_int; |
| 18 |
sub gen_bigint; |
| 19 |
sub gen_float; |
| 20 |
sub gen_char; |
| 21 |
sub gen_decimal; |
| 22 |
sub gen_bit; |
| 23 |
sub gen_year; |
| 24 |
sub gen_date; |
| 25 |
sub gen_time; |
| 26 |
sub gen_datetime; |
| 27 |
sub gen_timestamp; |
| 28 |
sub gen_blob; |
| 29 |
|
| 30 |
my %typeinfo = ( |
| 31 |
tinyint => { |
| 32 |
gen => \&gen_int, |
| 33 |
min_signed => -2**7, |
| 34 |
max_signed => 2**7-1, |
| 35 |
max_unsigned => 2**8-1, |
| 36 |
}, |
| 37 |
smallint => { |
| 38 |
gen => \&gen_int, |
| 39 |
min_signed => -2**15, |
| 40 |
max_signed => 2**15-1, |
| 41 |
max_unsigned => 2**16-1, |
| 42 |
}, |
| 43 |
mediumint => { |
| 44 |
gen => \&gen_int, |
| 45 |
min_signed => -2**23, |
| 46 |
max_signed => 2**23-1, |
| 47 |
max_unsigned => 2**24-1, |
| 48 |
}, |
| 49 |
int => { |
| 50 |
gen => \&gen_int, |
| 51 |
min_signed => -2**31, |
| 52 |
max_signed => 2**31-1, |
| 53 |
max_unsigned => 2**32-1, |
| 54 |
}, |
| 55 |
bigint => { |
| 56 |
gen => \&gen_bigint, |
| 57 |
min_signed => -2**31, |
| 58 |
max_signed => 2**31-1, |
| 59 |
max_unsigned => 2**32-1, |
| 60 |
}, |
| 61 |
double => { |
| 62 |
gen => \&gen_float, |
| 63 |
max_exp => 308, |
| 64 |
}, |
| 65 |
float => { |
| 66 |
gen => \&gen_float, |
| 67 |
max_exp => 38 |
| 68 |
}, |
| 69 |
char => { |
| 70 |
gen => \&gen_char, |
| 71 |
}, |
| 72 |
varchar => { |
| 73 |
gen => \&gen_char, |
| 74 |
}, |
| 75 |
binary => { |
| 76 |
binary => 1, |
| 77 |
gen => \&gen_char, |
| 78 |
}, |
| 79 |
varbinary => { |
| 80 |
binary => 1, |
| 81 |
gen => \&gen_char, |
| 82 |
}, |
| 83 |
decimal => { |
| 84 |
gen => \&gen_decimal, |
| 85 |
}, |
| 86 |
bit => { |
| 87 |
gen => \&gen_bit, |
| 88 |
}, |
| 89 |
year => { |
| 90 |
gen => \&gen_year, |
| 91 |
}, |
| 92 |
date => { |
| 93 |
gen => \&gen_date, |
| 94 |
}, |
| 95 |
time => { |
| 96 |
gen => \&gen_time, |
| 97 |
}, |
| 98 |
datetime => { |
| 99 |
gen => \&gen_datetime, |
| 100 |
}, |
| 101 |
timestamp => { |
| 102 |
gen => \&gen_timestamp, |
| 103 |
}, |
| 104 |
text => { |
| 105 |
gen => \&gen_blob, |
| 106 |
}, |
| 107 |
blob => { |
| 108 |
gen => \&gen_blob, |
| 109 |
}, |
| 110 |
); |
| 111 |
|
| 112 |
# init test |
| 113 |
|
| 114 |
# CSV control strings |
| 115 |
# note --exec uses sh so these are passed in single quotes |
| 116 |
# note using --csvopt now to avoid quoting (windows) |
| 117 |
|
| 118 |
sub get_csvfmt1 { |
| 119 |
my $csvfmt = { |
| 120 |
fields_terminated_by => '\t', |
| 121 |
fields_enclosed_by => undef, |
| 122 |
fields_escaped_by => '\\', |
| 123 |
lines_terminated_by => '\n', |
| 124 |
csvopt => 'n', # must match above |
| 125 |
}; |
| 126 |
return $csvfmt; |
| 127 |
} |
| 128 |
|
| 129 |
sub get_csvfmt2 { |
| 130 |
my $csvfmt = { |
| 131 |
fields_terminated_by => ',', |
| 132 |
fields_enclosed_by => '"', |
| 133 |
fields_escaped_by => '\\', |
| 134 |
lines_terminated_by => '\n', |
| 135 |
csvopt => 'cqn', # must match above |
| 136 |
}; |
| 137 |
return $csvfmt; |
| 138 |
} |
| 139 |
|
| 140 |
# translate used CSV control string to binary |
| 141 |
# in perl '\\' and '\'' are 1 byte and other '\x' are 2 bytes |
| 142 |
# basic double quoted chars work as expected |
| 143 |
|
| 144 |
sub translate_csv_control { |
| 145 |
my ($s) = @_; |
| 146 |
return "\t" if $s eq '\t'; |
| 147 |
return "\n" if $s eq '\n'; |
| 148 |
return $s if $s eq ','; |
| 149 |
return $s if $s eq '"'; |
| 150 |
return "\\" if $s eq '\\'; |
| 151 |
return "\\" if $s eq '\\\\'; |
| 152 |
die "translate_csv_control: unknown $s"; |
| 153 |
} |
| 154 |
|
| 155 |
sub init_test { |
| 156 |
my ($test) = @_; |
| 157 |
my $f = $test->{csvfmt}; |
| 158 |
my $t = $test->{csvesc} = {}; |
| 159 |
my @name = qw( |
| 160 |
fields_terminated_by |
| 161 |
fields_enclosed_by |
| 162 |
fields_escaped_by |
| 163 |
lines_terminated_by |
| 164 |
); |
| 165 |
for my $name (@name) { |
| 166 |
if (defined($f->{$name})) { |
| 167 |
$t->{$name} = translate_csv_control($f->{$name}); |
| 168 |
} else { |
| 169 |
$t->{$name} = undef; |
| 170 |
} |
| 171 |
} |
| 172 |
# defaults |
| 173 |
$test->{name} = "test->$test->{tag}"; |
| 174 |
$test->{statedir} = "$vardir/tmp"; |
| 175 |
$test->{csvdir} = "$vardir/tmp"; |
| 176 |
$test->{database} = "test"; |
| 177 |
if (defined($test->{verify})) { |
| 178 |
$test->{verify} =~ /^(all|pk)$/ |
| 179 |
or die "$test->{name}: bad verify option '$test->{verify}'"; |
| 180 |
} |
| 181 |
# does load data for verification need own csv file |
| 182 |
$test->{csvver} = $test->{verify} && $test->{rejectsgen}; |
| 183 |
# resume requires bad rows |
| 184 |
$test->{resumeopt} && !$test->{rejectsgen} |
| 185 |
and die "$test->{name}: resume requires rejected rows"; |
| 186 |
} |
| 187 |
|
| 188 |
sub get_tablename { |
| 189 |
my ($test, $table, $opts) = @_; |
| 190 |
my $tag = $test->{tag}; |
| 191 |
my $ver = $opts->{ver} ? "ver" : ""; |
| 192 |
return "$table->{name}$tag$ver"; |
| 193 |
} |
| 194 |
|
| 195 |
sub get_csvfile { |
| 196 |
my ($test, $table, $opts) = @_; |
| 197 |
my $tag = $test->{tag}; |
| 198 |
my $ver = $test->{csvver} && $opts->{ver} ? "ver" : ""; |
| 199 |
return "$test->{csvdir}/$table->{name}$tag$ver.csv"; |
| 200 |
} |
| 201 |
|
| 202 |
# create and drop |
| 203 |
|
| 204 |
sub create_attr { |
| 205 |
my ($attr) = @_; |
| 206 |
$typeinfo{$attr->{type}} |
| 207 |
or die "$attr->{name}: bad type: $attr->{type}"; |
| 208 |
my @txt = (); |
| 209 |
push(@txt, $attr->{name}); |
| 210 |
push(@txt, $attr->{type}); |
| 211 |
push(@txt, "($attr->{len})") |
| 212 |
if defined($attr->{len}); |
| 213 |
push(@txt, "($attr->{prec}, $attr->{scale})") |
| 214 |
if defined($attr->{prec}) && defined($attr->{scale}); |
| 215 |
push(@txt, "($attr->{prec})") |
| 216 |
if defined($attr->{prec}) && !defined($attr->{scale}); |
| 217 |
push(@txt, "unsigned") |
| 218 |
if $attr->{unsigned}; |
| 219 |
push(@txt, "not null") |
| 220 |
if $attr->{notnull}; |
| 221 |
return "@txt"; |
| 222 |
} |
| 223 |
|
| 224 |
sub create_attrs { |
| 225 |
my ($attrs) = @_; |
| 226 |
my @txt = (); |
| 227 |
my $attrno = 0; |
| 228 |
for my $attr (@$attrs) { |
| 229 |
$attr->{attrno} = $attrno++; |
| 230 |
$attr->{notnull} = 1 if $attr->{pk}; |
| 231 |
} |
| 232 |
for my $attr (@$attrs) { |
| 233 |
my $s = $attr->{attrno} == 0 ? "\n" : ",\n"; |
| 234 |
push(@txt, $s . create_attr($attr)); |
| 235 |
$attrno++; |
| 236 |
} |
| 237 |
return "@txt"; |
| 238 |
} |
| 239 |
|
| 240 |
sub create_table { |
| 241 |
my ($test, $table, $opts) = @_; |
| 242 |
my $tname = get_tablename($test, $table, $opts); |
| 243 |
my @txt = (); |
| 244 |
push(@txt, "create table $tname ("); |
| 245 |
my $attrs = $table->{attrs}; |
| 246 |
push(@txt, create_attrs($attrs)); |
| 247 |
my @pkattrs = (); |
| 248 |
for my $attr (@$attrs) { |
| 249 |
push(@pkattrs, $attr) |
| 250 |
if $attr->{pk}; |
| 251 |
} |
| 252 |
if (@pkattrs) { |
| 253 |
my $pklist = join(", ", map($_->{name}, @pkattrs)); |
| 254 |
push(@txt, ",\nprimary key ($pklist)"); |
| 255 |
} |
| 256 |
$table->{pkattrs} = [ @pkattrs ]; |
| 257 |
push(@txt, "\n) engine $opts->{engine};\n"); |
| 258 |
return "@txt"; |
| 259 |
} |
| 260 |
|
| 261 |
sub create_tables { |
| 262 |
my ($test, $opts) = @_; |
| 263 |
my @txt = (); |
| 264 |
my $tables = $test->{tables}; |
| 265 |
for my $table (@$tables) { |
| 266 |
push(@txt, create_table($test, $table, $opts)); |
| 267 |
} |
| 268 |
return "@txt"; |
| 269 |
}; |
| 270 |
|
| 271 |
sub drop_table { |
| 272 |
my ($test, $table, $opts) = @_; |
| 273 |
my $tname = get_tablename($test, $table, $opts); |
| 274 |
my @txt = (); |
| 275 |
push(@txt, "drop table"); |
| 276 |
push(@txt, "if exists") |
| 277 |
if $opts->{if_exists}; |
| 278 |
push(@txt, "$tname;\n"); |
| 279 |
return "@txt"; |
| 280 |
} |
| 281 |
|
| 282 |
sub drop_tables { |
| 283 |
my ($test, $opts) = @_; |
| 284 |
my @txt = (); |
| 285 |
my $tables = $test->{tables}; |
| 286 |
for my $table (@$tables) { |
| 287 |
push(@txt, drop_table($test, $table, $opts)); |
| 288 |
} |
| 289 |
return "@txt"; |
| 290 |
}; |
| 291 |
|
| 292 |
# run test |
| 293 |
|
| 294 |
sub run_import { |
| 295 |
my ($test, $opts) = @_; |
| 296 |
my $fter = $test->{csvfmt}{fields_terminated_by}; |
| 297 |
my $fenc = $test->{csvfmt}{fields_enclosed_by}; |
| 298 |
my @cmd = (); |
| 299 |
push(@cmd, "\$NDB_IMPORT"); |
| 300 |
push(@cmd, "--state-dir=$test->{statedir}"); |
| 301 |
push(@cmd, "--keep-state"); |
| 302 |
push(@cmd, "--stats"); |
| 303 |
push(@cmd, "--input-type=csv"); |
| 304 |
push(@cmd, "--output-type=ndb"); |
| 305 |
push(@cmd, "--temperrors=100"); |
| 306 |
# using line terminator "\n" |
| 307 |
push(@cmd, "--csvopt=n"); |
| 308 |
if (defined($test->{csvfmt}{csvopt})) { |
| 309 |
push(@cmd, "--csvopt=$test->{csvfmt}{csvopt}"); |
| 310 |
} else { |
| 311 |
push(@cmd, "--fields-terminated-by='$fter'"); |
| 312 |
if (defined($fenc)) { |
| 313 |
push(@cmd, "--fields-optionally-enclosed-by='$fenc'"); |
| 314 |
} |
| 315 |
} |
| 316 |
if ($test->{rejectsopt}) { |
| 317 |
push(@cmd, "--rejects=$test->{rejectsopt}"); |
| 318 |
} |
| 319 |
# $opts tells if this is a resume |
| 320 |
if ($opts->{resumeopt}) { |
| 321 |
push(@cmd, "--resume"); |
| 322 |
} |
| 323 |
push(@cmd, "--log-level=1"); |
| 324 |
push(@cmd, $test->{database}); |
| 325 |
my $tables = $test->{tables}; |
| 326 |
for my $table (@$tables) { |
| 327 |
my $csvfile = get_csvfile($test, $table, $opts); |
| 328 |
push(@cmd, $csvfile); |
| 329 |
} |
| 330 |
# runs mainly on unix/linux so stderr goes to stdout |
| 331 |
my $log = '>>$NDB_TOOLS_OUTPUT 2>&1'; |
| 332 |
my @cmd1 = ("--exec", "echo", @cmd, $log); |
| 333 |
my @cmd2 = ("--exec", @cmd, $log); |
| 334 |
if (defined($opts->{error})) { |
| 335 |
unshift(@cmd2, "--error $opts->{error}\n"); |
| 336 |
} |
| 337 |
return "@cmd1\n@cmd2\n"; |
| 338 |
} |
| 339 |
|
| 340 |
sub select_count { |
| 341 |
my ($test, $table, $opts) = @_; |
| 342 |
my $tname = get_tablename($test, $table, $opts); |
| 343 |
my @txt = (); |
| 344 |
push(@txt, "select count(*) from $tname;\n"); |
| 345 |
if ($test->{dumpdata}) { |
| 346 |
my $file = "$vardir/tmp/$tname.dump"; |
| 347 |
push @txt, "--disable_query_log\n"; |
| 348 |
push(@txt, "select * from $tname order by 1\ninto outfile '$file';\n"); |
| 349 |
push @txt, "--enable_query_log\n"; |
| 350 |
} |
| 351 |
return "@txt"; |
| 352 |
} |
| 353 |
|
| 354 |
sub select_counts { |
| 355 |
my ($test, $opts) = @_; |
| 356 |
my @txt = (); |
| 357 |
my $tables = $test->{tables}; |
| 358 |
for my $table (@$tables) { |
| 359 |
push(@txt, select_count($test, $table, $opts)); |
| 360 |
} |
| 361 |
return "@txt"; |
| 362 |
} |
| 363 |
|
| 364 |
sub load_table { |
| 365 |
my ($test, $table, $opts) = @_; |
| 366 |
my $csvfile = get_csvfile($test, $table, $opts); |
| 367 |
my $tname = get_tablename($test, $table, $opts); |
| 368 |
my $fter = $test->{csvfmt}{fields_terminated_by}; |
| 369 |
my $fenc = $test->{csvfmt}{fields_enclosed_by}; |
| 370 |
my @txt = (); |
| 371 |
push(@txt, "load data infile '$csvfile'\n"); |
| 372 |
push(@txt, "into table $tname\n"); |
| 373 |
# we generate 8-bit ascii so use some charset accepting it |
| 374 |
push(@txt, "character set latin1\n"); |
| 375 |
push(@txt, "fields"); |
| 376 |
push(@txt, "terminated by '$fter'"); |
| 377 |
if (defined($fenc)) { |
| 378 |
push(@txt, "optionally enclosed by '$fenc'"); |
| 379 |
} |
| 380 |
push(@txt, "lines terminated by '\\n'"); |
| 381 |
push(@txt, ";\n"); |
| 382 |
return "@txt"; |
| 383 |
} |
| 384 |
|
| 385 |
sub load_tables { |
| 386 |
my ($test, $opts) = @_; |
| 387 |
my @txt = (); |
| 388 |
push @txt, "--disable_query_log\n"; |
| 389 |
my $tables = $test->{tables}; |
| 390 |
for my $table (@$tables) { |
| 391 |
push(@txt, load_table($test, $table, $opts)); |
| 392 |
} |
| 393 |
push @txt, "--enable_query_log\n"; |
| 394 |
return "@txt"; |
| 395 |
} |
| 396 |
|
| 397 |
sub verify_table { |
| 398 |
my ($test, $table, $opts) = @_; |
| 399 |
my $tname1 = get_tablename($test, $table, { ver => 0 }); |
| 400 |
my $tname2 = get_tablename($test, $table, { ver => 1 }); |
| 401 |
my @txt = (); |
| 402 |
push(@txt, "select count(*) from $tname1 x, $tname2 y\n"); |
| 403 |
push(@txt, "where"); |
| 404 |
my @cls = (); |
| 405 |
my $attrs = $table->{attrs}; |
| 406 |
for my $attr (@$attrs) { |
| 407 |
if ($test->{verify} eq 'pk' && !$attr->{pk}) { |
| 408 |
next; |
| 409 |
} |
| 410 |
my $a1 = "x." . $attr->{name}; |
| 411 |
my $a2 = "y." . $attr->{name}; |
| 412 |
my $c; |
| 413 |
if ($attr->{notnull}) { |
| 414 |
$c = "$a1 = $a2"; |
| 415 |
} else { |
| 416 |
$c = "($a1 = $a2 or ($a1 is null and $a2 is null))"; |
| 417 |
} |
| 418 |
push(@cls, $c); |
| 419 |
} |
| 420 |
push(@txt, join(" and\n", @cls)); |
| 421 |
push(@txt, ";\n"); |
| 422 |
return "@txt"; |
| 423 |
} |
| 424 |
|
| 425 |
sub verify_tables { |
| 426 |
my ($test, $opts) = @_; |
| 427 |
my @txt = (); |
| 428 |
my $tables = $test->{tables}; |
| 429 |
for my $table (@$tables) { |
| 430 |
push(@txt, verify_table($test, $table, $opts)); |
| 431 |
} |
| 432 |
return "@txt"; |
| 433 |
} |
| 434 |
|
| 435 |
# write test |
| 436 |
|
| 437 |
sub make_test { |
| 438 |
my ($test) = @_; |
| 439 |
init_test($test); |
| 440 |
my @txt = (); |
| 441 |
push @txt, "--echo # test $test->{tag} - $test->{desc}\n"; |
| 442 |
push @txt, create_tables($test, { engine => "ndb" }); |
| 443 |
if (!$test->{resumeopt}) { |
| 444 |
push @txt, run_import($test, {}); |
| 445 |
} else { |
| 446 |
push @txt, run_import($test, { error => "1" }); |
| 447 |
for (my $i = 1; $i <= $test->{rejectsgen}; $i++) { |
| 448 |
push @txt, run_import($test, { resumeopt => $i, error => "0,1" }); |
| 449 |
} |
| 450 |
} |
| 451 |
push @txt, select_counts($test, {}); |
| 452 |
if ($test->{verify}) { |
| 453 |
push @txt, create_tables($test, { engine => "ndb", ver => 1 }); |
| 454 |
push @txt, load_tables($test, { ver => 1 }); |
| 455 |
push @txt, select_counts($test, { ver => 1 }); |
| 456 |
push @txt, verify_tables($test, {}); |
| 457 |
push @txt, drop_tables($test, { ver => 1 }); |
| 458 |
} |
| 459 |
push @txt, drop_tables($test, {}); |
| 460 |
return "@txt"; |
| 461 |
} |
| 462 |
|
| 463 |
sub write_tests { |
| 464 |
my ($tests) = @_; |
| 465 |
my $tag = $tests->{tag}; |
| 466 |
my $file = "$vardir/tmp/ndb_import$tag.inc"; |
| 467 |
my $fh = gensym(); |
| 468 |
open($fh, ">:raw", $file) |
| 469 |
or die "$file: open for write failed: $!"; |
| 470 |
my $testlist = $tests->{testlist}; |
| 471 |
for my $test (@$testlist) { |
| 472 |
print $fh make_test($test); |
| 473 |
} |
| 474 |
close($fh) |
| 475 |
or die "$file: close after write failed: $!"; |
| 476 |
for my $test (@$testlist) { |
| 477 |
# values in $opts propagate down *and* up |
| 478 |
my $opts = {}; |
| 479 |
write_csvfiles($test, $opts); |
| 480 |
} |
| 481 |
} |
| 482 |
|
| 483 |
# generate values |
| 484 |
|
| 485 |
sub myrand { |
| 486 |
my ($m) = @_; |
| 487 |
return int(rand($m)); |
| 488 |
} |
| 489 |
|
| 490 |
sub myrand2 { |
| 491 |
my ($m, $k) = @_; |
| 492 |
my $n = myrand($m); |
| 493 |
while ($k > 0) { |
| 494 |
$n = myrand($n + 1); |
| 495 |
$k--; |
| 496 |
} |
| 497 |
return $n; |
| 498 |
} |
| 499 |
|
| 500 |
sub gen_int { |
| 501 |
my ($test, $attr, $opts) = @_; |
| 502 |
my $typeinfo = $typeinfo{$attr->{type}}; |
| 503 |
my $lo; |
| 504 |
my $hi; |
| 505 |
if ($attr->{unsigned}) { |
| 506 |
$lo = 0; |
| 507 |
$hi = $typeinfo->{max_unsigned}; |
| 508 |
} else { |
| 509 |
$lo = $typeinfo->{min_signed}; |
| 510 |
$hi = $typeinfo->{max_signed}; |
| 511 |
} |
| 512 |
my $val; |
| 513 |
if ($attr->{pk}) { |
| 514 |
$val = $opts->{rowid} % ($hi + 1); |
| 515 |
} else { |
| 516 |
my $r = myrand(100); |
| 517 |
if ($r < $par_sppct) { |
| 518 |
$val = 0 if $r % 3 == 0; |
| 519 |
$val = $lo if $r % 3 == 1; |
| 520 |
$val = $hi if $r % 3 == 2; |
| 521 |
} else { |
| 522 |
$val = $lo + myrand($hi - $lo + 1); |
| 523 |
} |
| 524 |
} |
| 525 |
if ($attr->{type} eq 'smallint' && |
| 526 |
$opts->{rejectsflag} && |
| 527 |
myrand(100) < $par_valerrpct) { |
| 528 |
if ($attr->{unsigned}) { |
| 529 |
$val = $typeinfo->{max_unsigned} + 1; |
| 530 |
} elsif (myrand(2) == 0) { |
| 531 |
$val = $typeinfo->{max_signed} + 1; |
| 532 |
} else { |
| 533 |
$val = $typeinfo->{min_signed} - 1; |
| 534 |
} |
| 535 |
$opts->{rejectserrs} .= ":value"; |
| 536 |
} |
| 537 |
return "$val"; |
| 538 |
} |
| 539 |
|
| 540 |
sub gen_bigint { |
| 541 |
my ($test, $attr, $opts) = @_; |
| 542 |
my $val = gen_int($test, $attr, $opts); |
| 543 |
return $val; |
| 544 |
}; |
| 545 |
|
| 546 |
sub gen_float { |
| 547 |
my ($test, $attr, $opts) = @_; |
| 548 |
my $typeinfo = $typeinfo{$attr->{type}}; |
| 549 |
my $hiexp = $typeinfo->{max_exp}; |
| 550 |
my $p = myrand2($hiexp + 1, $par_randbias); |
| 551 |
if (myrand(2) == 0) { |
| 552 |
$p = (-1) * $p; |
| 553 |
} |
| 554 |
# floating rand |
| 555 |
my $v = rand() * (10 ** $p); |
| 556 |
if (myrand(2) == 0) { |
| 557 |
$v = (-1) * $v; |
| 558 |
} |
| 559 |
my $val; |
| 560 |
if (myrand(2) == 0) { |
| 561 |
$val = sprintf("%f", $v); |
| 562 |
} else { |
| 563 |
$val = sprintf("%g", $v); |
| 564 |
} |
| 565 |
if ($attr->{type} eq 'float' && |
| 566 |
$opts->{rejectsflag} && |
| 567 |
myrand(100) < $par_valerrpct) { |
| 568 |
$val = sprintf("1e%d", $hiexp + 10); |
| 569 |
$opts->{rejectserrs} .= ":value"; |
| 570 |
} |
| 571 |
return $val; |
| 572 |
} |
| 573 |
|
| 574 |
my $g_escape = { |
| 575 |
0 => '0', |
| 576 |
0x08 => 'b', |
| 577 |
0x0a => 'n', |
| 578 |
0x0d => 'r', |
| 579 |
0x09 => 't', |
| 580 |
0x1a => 'Z', |
| 581 |
}; |
| 582 |
|
| 583 |
sub make_byte { |
| 584 |
my ($test, $opts) = @_; |
| 585 |
my $fter = $test->{csvesc}{fields_terminated_by}; |
| 586 |
my $fenc = $test->{csvesc}{fields_enclosed_by}; |
| 587 |
my $fesc = $test->{csvesc}{fields_escaped_by}; |
| 588 |
my $lter = $test->{csvesc}{lines_terminated_by}; |
| 589 |
my $binary = $opts->{binary}; |
| 590 |
my $mask = $opts->{mask}; |
| 591 |
my $val; |
| 592 |
while (1) { |
| 593 |
my $x; |
| 594 |
# non-binary used only to make readable files, no charsets yet |
| 595 |
if (!$binary) { |
| 596 |
$x = 0x61 + myrand(26); |
| 597 |
} else { |
| 598 |
$x = myrand(256); |
| 599 |
} |
| 600 |
if (defined($mask)) { |
| 601 |
$x &= $mask; |
| 602 |
} |
| 603 |
if ($x == 0) { |
| 604 |
$val = $fesc.'0'; |
| 605 |
last; |
| 606 |
} |
| 607 |
if (0 && $x == 032) { # now using _O_BINARY on windows |
| 608 |
$val = $fesc.'Z'; |
| 609 |
last; |
| 610 |
} |
| 611 |
if ($x == ord($fter)) { |
| 612 |
if ($opts->{quote}) { |
| 613 |
$val = $fter; |
| 614 |
} else { |
| 615 |
$val = $fesc.$fter; |
| 616 |
} |
| 617 |
last; |
| 618 |
} |
| 619 |
if (defined($fenc) && $x == ord($fenc)) { |
| 620 |
if ($opts->{quote} && myrand(2) == 0) { |
| 621 |
$val = $fenc.$fenc; |
| 622 |
} else { |
| 623 |
$val = $fesc.$fenc; |
| 624 |
} |
| 625 |
last; |
| 626 |
} |
| 627 |
if ($x == ord($fesc)) { |
| 628 |
$val = $fesc.$fesc; |
| 629 |
last; |
| 630 |
} |
| 631 |
if ($x == ord($lter)) { |
| 632 |
if ($opts->{quote}) { |
| 633 |
if (myrand(2) == 0) { |
| 634 |
$val = $lter; |
| 635 |
last; |
| 636 |
} |
| 637 |
} |
| 638 |
if ($lter eq "\n") { |
| 639 |
$val = "\\n"; |
| 640 |
last; |
| 641 |
} |
| 642 |
die "make_byte: cannot handle lter='$lter'"; |
| 643 |
} |
| 644 |
if (myrand(5) == 0) { |
| 645 |
my $v = $g_escape->{$x}; |
| 646 |
if (defined($v)) { |
| 647 |
$val = $fesc.$v; |
| 648 |
last; |
| 649 |
} |
| 650 |
} |
| 651 |
$val = pack('C', $x); |
| 652 |
last; |
| 653 |
} |
| 654 |
return $val; |
| 655 |
} |
| 656 |
|
| 657 |
sub make_string { |
| 658 |
my ($test, $opts) = @_; |
| 659 |
my $hilen = $opts->{hilen}; |
| 660 |
my $len = myrand(10) != 0 ? myrand2($hilen, $par_randbias) : $hilen; |
| 661 |
if ($len == 0 && $opts->{minlen}) { |
| 662 |
$len = $opts->{minlen}; |
| 663 |
} |
| 664 |
my $val = ""; |
| 665 |
for (my $i = 0; $i < $len; $i++) { |
| 666 |
$val .= make_byte($test, $opts); |
| 667 |
} |
| 668 |
return $val; |
| 669 |
} |
| 670 |
|
| 671 |
sub gen_char { |
| 672 |
my ($test, $attr, $opts) = @_; |
| 673 |
my $typeinfo = $typeinfo{$attr->{type}}; |
| 674 |
my $val; |
| 675 |
if ($attr->{pk}) { |
| 676 |
$val = sprintf("%d", $opts->{rowid}); |
| 677 |
} else { |
| 678 |
$opts->{binary} = $typeinfo->{binary}; |
| 679 |
$opts->{hilen} = $attr->{len}; |
| 680 |
$opts->{minlen} = $attr->{minlen}; |
| 681 |
$val = make_string($test, $opts); |
| 682 |
} |
| 683 |
return $val; |
| 684 |
} |
| 685 |
|
| 686 |
sub gen_bit { |
| 687 |
my ($test, $attr, $opts) = @_; |
| 688 |
my $typeinfo = $typeinfo{$attr->{type}}; |
| 689 |
my $fter = $test->{csvfmt}{fields_terminated_by}; |
| 690 |
my $fesc = $test->{csvfmt}{fields_escaped_by}; |
| 691 |
my $l = $attr->{len}; |
| 692 |
my $n = int(($l + 7) / 8); # bytes rounded up |
| 693 |
my $m = 8 - (8 * $n - $l); # bits in last byte |
| 694 |
my $val = ""; |
| 695 |
for (my $i = 0; $i < $n; $i++) { |
| 696 |
my $mask = 255; |
| 697 |
if ($i + 1 == $n && $m != 0) { |
| 698 |
$mask = (1 << $m) - 1; |
| 699 |
} |
| 700 |
$opts->{mask} = $mask; |
| 701 |
my $v = make_byte($test, $opts); |
| 702 |
$val .= $v; |
| 703 |
} |
| 704 |
return $val; |
| 705 |
} |
| 706 |
|
| 707 |
sub make_decimalpart { |
| 708 |
my ($test, $opts) = @_; |
| 709 |
my $n = $opts->{partlen}; |
| 710 |
my $m = myrand($n + 1); |
| 711 |
my $val = ""; |
| 712 |
for (my $k = 0; $k < $m; $k++) { |
| 713 |
my $d = myrand(10); |
| 714 |
$val .= sprintf("%d", $d); |
| 715 |
} |
| 716 |
return $val; |
| 717 |
} |
| 718 |
|
| 719 |
sub gen_decimal { |
| 720 |
my ($test, $attr, $opts) = @_; |
| 721 |
my $prec = $attr->{prec}; |
| 722 |
my $scale = $attr->{scale}; |
| 723 |
$prec >= $scale or die "invalid prec=$prec scale=$scale"; |
| 724 |
my $val; |
| 725 |
while (1) { |
| 726 |
$opts->{partlen} = $prec - $scale; |
| 727 |
$val = make_decimalpart($test, $opts); |
| 728 |
if (myrand(2) == 0) { |
| 729 |
$val .= "."; |
| 730 |
if (myrand(2) == 0) { |
| 731 |
$opts->{partlen} = $scale; |
| 732 |
$val .= make_decimalpart($test, $opts); |
| 733 |
} |
| 734 |
} |
| 735 |
last if $val =~ /\d/; |
| 736 |
} |
| 737 |
if ($attr->{unsigned}) { |
| 738 |
if (myrand(3) == 0) { |
| 739 |
$val = "+$val"; |
| 740 |
} |
| 741 |
} else { |
| 742 |
if (myrand(3) == 0) { |
| 743 |
$val = "-$val"; |
| 744 |
} |
| 745 |
} |
| 746 |
return $val; |
| 747 |
} |
| 748 |
|
| 749 |
sub gen_year { |
| 750 |
my ($test, $attr, $opts) = @_; |
| 751 |
my $yy = 1901 + myrand(255); |
| 752 |
my $val = sprintf("%04d", $yy); |
| 753 |
return $val; |
| 754 |
} |
| 755 |
|
| 756 |
# Date::Manip could be used to generate better valid dates |
| 757 |
sub make_date { |
| 758 |
my ($test, $opts) = @_; |
| 759 |
my $yystart = $opts->{yystart}; |
| 760 |
my $yyrange = $opts->{yyrange}; |
| 761 |
my $yy = $yystart + myrand($yyrange); |
| 762 |
my $mm = 1 + myrand(12); |
| 763 |
my $dd = 1 + myrand(28); |
| 764 |
my $val = sprintf("%04d-%02d-%02d", $yy, $mm, $dd); |
| 765 |
return $val; |
| 766 |
} |
| 767 |
|
| 768 |
sub gen_date { |
| 769 |
my ($test, $attr, $opts) = @_; |
| 770 |
$opts->{yystart} = 1950; |
| 771 |
$opts->{yyrange} = 100; |
| 772 |
my $val = make_date($test, $opts); |
| 773 |
return $val; |
| 774 |
} |
| 775 |
|
| 776 |
sub make_frac { |
| 777 |
my ($test, $opts) = @_; |
| 778 |
my $prec = $opts->{prec}; |
| 779 |
my $n = myrand($prec + 1); |
| 780 |
my $val = ""; |
| 781 |
for (my $i = 0; $i < $n; $i++) { |
| 782 |
$val .= myrand(10); |
| 783 |
} |
| 784 |
$val = ".$val" if $n != 0 || myrand(5) == 0; |
| 785 |
return $val; |
| 786 |
} |
| 787 |
|
| 788 |
sub make_time { |
| 789 |
my ($test, $opts) = @_; |
| 790 |
my $sep = $opts->{sep}; |
| 791 |
my $ts = $opts->{ts}; |
| 792 |
my $val; |
| 793 |
my $hh = myrand(24); |
| 794 |
if ($ts && ($hh == 2 || $hh == 3 || $hh == 4)) { |
| 795 |
$hh = 1; |
| 796 |
} |
| 797 |
my $mm = myrand(60); |
| 798 |
my $ss = myrand(60); |
| 799 |
if ($sep) { |
| 800 |
$hh = "0$hh" if $hh < 10 && myrand(2) == 0; |
| 801 |
$mm = "0$mm" if $mm < 10 && myrand(2) == 0; |
| 802 |
$ss = "0$ss" if $ss < 10 && myrand(2) == 0; |
| 803 |
$val = "$hh:$mm:$ss"; |
| 804 |
} else { |
| 805 |
$hh = "0$hh" if $hh < 10; |
| 806 |
$mm = "0$mm" if $mm < 10; |
| 807 |
$ss = "0$ss" if $ss < 10; |
| 808 |
$val = "$hh$mm$ss"; |
| 809 |
} |
| 810 |
$val .= make_frac($test, $opts); |
| 811 |
return $val; |
| 812 |
} |
| 813 |
|
| 814 |
sub gen_time { |
| 815 |
my ($test, $attr, $opts) = @_; |
| 816 |
$opts->{sep} = myrand(2); |
| 817 |
$opts->{prec} = $attr->{prec}; |
| 818 |
$opts->{ts} = 0; |
| 819 |
my $val = make_time($test, $opts); |
| 820 |
return $val; |
| 821 |
} |
| 822 |
|
| 823 |
sub gen_datetime { |
| 824 |
my ($test, $attr, $opts) = @_; |
| 825 |
$opts->{yystart} = 1950; |
| 826 |
$opts->{yyrange} = 100; |
| 827 |
my $d = make_date($test, $opts); |
| 828 |
$opts->{sep} = 1; |
| 829 |
$opts->{prec} = $attr->{prec}; |
| 830 |
$opts->{ts} = 0; |
| 831 |
my $t = make_time($test, $opts); |
| 832 |
my $p = " "; |
| 833 |
return "$d$p$t" |
| 834 |
} |
| 835 |
|
| 836 |
sub gen_timestamp { |
| 837 |
my ($test, $attr, $opts) = @_; |
| 838 |
my $prec = $attr->{prec}; |
| 839 |
$opts->{yystart} = 1971; |
| 840 |
$opts->{yyrange} = 64; |
| 841 |
my $d = make_date($test, $opts); |
| 842 |
$opts->{sep} = 1; |
| 843 |
$opts->{prec} = $attr->{prec}; |
| 844 |
$opts->{ts} = 1; |
| 845 |
my $t = make_time($test, $opts); |
| 846 |
my $p = " "; |
| 847 |
return "$d$p$t" |
| 848 |
} |
| 849 |
|
| 850 |
sub gen_blob { |
| 851 |
my ($test, $attr, $opts) = @_; |
| 852 |
my $typeinfo = $typeinfo{$attr->{type}}; |
| 853 |
$opts->{binary} = $typeinfo->{binary}; |
| 854 |
$opts->{hilen} = $attr->{len}; |
| 855 |
my $val = make_string($test, $opts); |
| 856 |
return $val; |
| 857 |
} |
| 858 |
|
| 859 |
sub gen_csvfield { |
| 860 |
my ($test, $table, $attr, $opts) = @_; |
| 861 |
my $typeinfo = $typeinfo{$attr->{type}}; |
| 862 |
my $fenc = $test->{csvesc}{fields_enclosed_by}; |
| 863 |
my $fesc = $test->{csvesc}{fields_escaped_by}; |
| 864 |
my $val; |
| 865 |
# if value will be enclosed by |
| 866 |
$opts->{quote} = defined($fenc) && |
| 867 |
myrand(100) < $par_quotepct; |
| 868 |
if (!$attr->{notnull} && |
| 869 |
myrand(100) < $par_nullpct) { |
| 870 |
$val = $fesc.'N'; |
| 871 |
} else { |
| 872 |
$val = $typeinfo->{gen}($test, $attr, $opts); |
| 873 |
if ($opts->{rejectsflag} && |
| 874 |
myrand(100) < $par_nullpct) { |
| 875 |
$val = $fesc.'N'; |
| 876 |
$opts->{rejectserrs} .= ":null"; |
| 877 |
} |
| 878 |
} |
| 879 |
if ($opts->{quote}) { |
| 880 |
$val = "$fenc$val$fenc"; |
| 881 |
} |
| 882 |
return $val; |
| 883 |
} |
| 884 |
|
| 885 |
sub get_pkvals { |
| 886 |
my ($table, $line) = @_; |
| 887 |
my $pkattrs = $table->{pkattrs}; |
| 888 |
my @pkvals = (); |
| 889 |
for my $attr (@$pkattrs) { |
| 890 |
my $n = $attr->{attrno}; |
| 891 |
push(@pkvals, $line->[$n]); |
| 892 |
} |
| 893 |
return [ @pkvals ]; |
| 894 |
} |
| 895 |
|
| 896 |
sub set_pkvals { |
| 897 |
my ($table, $line, $pkvals) = @_; |
| 898 |
my $pkattrs = $table->{pkattrs}; |
| 899 |
my $k = 0; |
| 900 |
for my $attr (@$pkattrs) { |
| 901 |
my $n = $attr->{attrno}; |
| 902 |
$line->[$n] = $pkvals->[$k]; |
| 903 |
$k++; |
| 904 |
} |
| 905 |
} |
| 906 |
|
| 907 |
sub gen_csvline { |
| 908 |
my ($test, $table, $opts) = @_; |
| 909 |
my $tname = get_tablename($test, $table, { ver => 0 }); |
| 910 |
my $attrs = $table->{attrs}; |
| 911 |
my $rowid = $opts->{rowid}; |
| 912 |
my $fter = $test->{csvesc}{fields_terminated_by}; |
| 913 |
my $fesc = $test->{csvesc}{fields_escaped_by}; |
| 914 |
my $lter = $test->{csvesc}{lines_terminated_by}; |
| 915 |
my @line = (); |
| 916 |
# per line |
| 917 |
$opts->{rejectserrs} = ""; |
| 918 |
for my $attr (@$attrs) { |
| 919 |
my $val = gen_csvfield($test, $table, $attr, $opts); |
| 920 |
push(@line, $val); |
| 921 |
} |
| 922 |
# if no errors from fields, create error on the row |
| 923 |
if ($opts->{rejectsflag}) { |
| 924 |
while ($opts->{rejectserrs} eq "") { |
| 925 |
if (myrand(2) == 0 && $rowid > 0) { |
| 926 |
# duplicate pk (one will be accepted, could be this one) |
| 927 |
my $oldrowid = myrand($rowid); |
| 928 |
my $oldpkvals = $opts->{pkvals}{$oldrowid}; |
| 929 |
if (defined($oldpkvals)) { |
| 930 |
set_pkvals($table, \@line, $oldpkvals); |
| 931 |
$opts->{rejectserrs} .= ":dubpk-$oldrowid"; |
| 932 |
last; |
| 933 |
} |
| 934 |
} |
| 935 |
if (myrand(2) == 0) { |
| 936 |
# too many fields |
| 937 |
my $val = $fesc.'N'; |
| 938 |
push(@line, $val); |
| 939 |
$opts->{rejectserrs} .= ":fields+1"; |
| 940 |
last; |
| 941 |
} |
| 942 |
if (myrand(2) == 0) { |
| 943 |
# too few fields |
| 944 |
pop(@line); |
| 945 |
$opts->{rejectserrs} .= ":fields-1"; |
| 946 |
last; |
| 947 |
} |
| 948 |
} |
| 949 |
} |
| 950 |
# save generated rejects |
| 951 |
($opts->{rejectsflag} == ($opts->{rejectserrs} ne "")) |
| 952 |
or die "rejects flag=$opts->{rejectsflag} errs=$opts->{rejectserrs}"; |
| 953 |
if ($opts->{rejectsflag}) { |
| 954 |
$opts->{rejectslines}{$rowid} = [ @line, $opts->{rejectserrs} ]; |
| 955 |
} |
| 956 |
# save pk values if line not rejected |
| 957 |
if (!$opts->{rejectsflag}) |
| 958 |
{ |
| 959 |
my $pkvals = get_pkvals($table, \@line); |
| 960 |
$opts->{pkvals}{$rowid} = $pkvals; |
| 961 |
} |
| 962 |
my $line = join($fter, @line).$lter; |
| 963 |
return $line; |
| 964 |
} |
| 965 |
|
| 966 |
# write csv files |
| 967 |
|
| 968 |
sub write_csvfile { |
| 969 |
my ($test, $table, $opts) = @_; |
| 970 |
my $file = get_csvfile($test, $table, { ver => 0 }); |
| 971 |
my $filever = get_csvfile($test, $table, { ver => 1 }); |
| 972 |
my $fh = gensym(); |
| 973 |
my $fhver = gensym(); |
| 974 |
open($fh, ">:raw", $file) |
| 975 |
or die "$file: open for write failed: $!"; |
| 976 |
if ($test->{csvver}) { |
| 977 |
open($fhver, ">:raw", $filever) |
| 978 |
or die "$filever: open for write failed: $!"; |
| 979 |
} |
| 980 |
my $tname = get_tablename($test, $table, { ver => 0 }); |
| 981 |
my $rows = $table->{rows}; |
| 982 |
# per file |
| 983 |
$opts->{pkvals} = {}; |
| 984 |
$opts->{rejectscnt} = 0; |
| 985 |
$opts->{rejectslines} = {}; |
| 986 |
for (my $n = 0; $n < $rows; $n++) { |
| 987 |
$opts->{rowid} = $n; |
| 988 |
$opts->{rejectsflag} = 0; |
| 989 |
if ($test->{rejectsgen}) { |
| 990 |
my $rowsleft = $rows - $n; |
| 991 |
my $rejectsleft = $test->{rejectsgen} - $opts->{rejectscnt}; |
| 992 |
if ($rejectsleft != 0) { |
| 993 |
if ($rejectsleft == $rowsleft || |
| 994 |
myrand(1 + $rowsleft/$rejectsleft) == 0) { |
| 995 |
$opts->{rejectsflag} = 1; |
| 996 |
$opts->{rejectserrs} = 0; |
| 997 |
} |
| 998 |
} |
| 999 |
} |
| 1000 |
my $line = gen_csvline($test, $table, $opts); |
| 1001 |
if ($opts->{rejectsflag}) { |
| 1002 |
$opts->{rejectserrs} or die "no rejectserrs"; |
| 1003 |
$opts->{rejectscnt}++; |
| 1004 |
} |
| 1005 |
print $fh "$line" |
| 1006 |
or die "$file: rowid $n: write failed: $!"; |
| 1007 |
if ($test->{csvver}) { |
| 1008 |
if (!$opts->{rejectsflag}) { |
| 1009 |
print $fhver "$line" |
| 1010 |
or die "$filever: rowid $n: write failed: $!"; |
| 1011 |
} |
| 1012 |
} |
| 1013 |
} |
| 1014 |
if ($test->{rejectsgen}) { |
| 1015 |
write_rejects($test, $table, $opts); |
| 1016 |
} |
| 1017 |
$test->{rejectsgen} == $opts->{rejectscnt} |
| 1018 |
or die "$tname: $test->{rejectsgen} != $opts->{rejectscnt}"; |
| 1019 |
close($fh) |
| 1020 |
or die "$file: close after write failed: $!"; |
| 1021 |
if ($test->{csvver}) { |
| 1022 |
close($fhver) |
| 1023 |
or die "$filever: close after write failed: $!"; |
| 1024 |
} |
| 1025 |
} |
| 1026 |
|
| 1027 |
# debug |
| 1028 |
sub write_rejects { |
| 1029 |
my ($test, $table, $opts) = @_; |
| 1030 |
my $tname = get_tablename($test, $table, { ver => 0 }); |
| 1031 |
my $file = "$vardir/tmp/$tname.genrej"; |
| 1032 |
my $fh = gensym(); |
| 1033 |
open($fh, ">:raw", $file) |
| 1034 |
or die "$file: open for write failed: $!"; |
| 1035 |
my $lines = $opts->{rejectslines}; |
| 1036 |
my @rowid = sort { $a <=> $b } keys %$lines; |
| 1037 |
for my $rowid (@rowid) { |
| 1038 |
my $line = $lines->{$rowid}; |
| 1039 |
my @out = join("\t", $rowid, @$line); |
| 1040 |
print $fh @out, "\n"; |
| 1041 |
} |
| 1042 |
close($fh) |
| 1043 |
or die "$file: close after write failed: $!"; |
| 1044 |
} |
| 1045 |
|
| 1046 |
sub write_csvfiles { |
| 1047 |
my ($test, $opts) = @_; |
| 1048 |
my $tables = $test->{tables}; |
| 1049 |
for my $table (@$tables) { |
| 1050 |
write_csvfile($test, $table, $opts); |
| 1051 |
} |
| 1052 |
} |
| 1053 |
|
| 1054 |
1; |