percona/percona-toolkit
1
0
Fork 0
mirror of https://github.com/percona/percona-toolkit.git synced 2025-09-04 03:26:19 +00:00
Code Issues Packages Projects Releases Wiki Activity

Rename preserve_embedded_numbers to match_embedded_numbers, and fingerprint_md5 to match_md5_checksums. Add corresponding options to pt-fingerprint.

Browse Source
This commit is contained in:
Daniel Nichter
2012-03-26 16:40:46 -06:00
parent a81d25c489
commit beaa9240e7
3 changed files with 51 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

45
bin/pt-fingerprint
View File

@@ -1559,8 +1559,25 @@ sub fingerprint {
$query =~ s/\\["']//g; # quoted strings
$query =~ s/".*?"/?/sg; # quoted strings
$query =~ s/'.*?'/?/sg; # quoted strings
$query =~ s/[0-9+-][0-9a-f.xb+-]*/?/g;# Anything vaguely resembling numbers
$query =~ s/[xb.+-]\?/?/g; # Clean up leftovers
if ( $self->{match_md5_checksums} ) {
$query =~ s/([._-])[a-f0-9]{32}/$1?/g;
}
if ( !$self->{match_embedded_numbers} ) {
$query =~ s/[0-9+-][0-9a-f.xb+-]*/?/g;
}
else {
$query =~ s/\b[0-9+-][0-9a-f.xb+-]*/?/g;
}
if ( $self->{match_md5_checksums} ) {
$query =~ s/[xb+-]\?/?/g;
}
else {
$query =~ s/[xb.+-]\?/?/g;
}
$query =~ s/\A\s+//; # Chop off leading whitespace
chomp $query; # Kill trailing whitespace
$query =~ tr[ \n\t\r\f][ ]s; # Collapse whitespace
@@ -1840,7 +1857,11 @@ sub main {
$o->usage_or_errors();
my $qp = new QueryParser();
my $qr = new QueryRewriter(QueryParser=>$qp);
my $qr = new QueryRewriter(
QueryParser => $qp,
match_md5_checksums => $o->get('match-md5-checksums'),
match_embedded_numbers => $o->get('match-embedded-numbers'),
);
if ( $o->got('query') ) {
print $qr->fingerprint($o->get('query')), "\n";
@@ -1993,6 +2014,24 @@ first option on the command line.
Show help and exit.
=item --match-embedded-numbers
Match numbers embedded in words and replace as single values. This option
causes the tool to be more careful about matching numbers so that words
with numbers, like C<catch22> are matched and replaced as a single C<?>
placeholder. Otherwise the default number matching pattern will replace
C<catch22> as C<catch?>.
This is helpful if database or table names contain numbers.
=item --match-md5-checksums
Match MD5 checksums and replace as single values. This option causes
the tool to be more careful about matching numbers so that MD5 checksums
like C<fbc5e685a5d3d45aa1d0347fdb7c4d35> are matched and replaced as a
single C<?> placeholder. Otherwise, the default number matching pattern will
replace C<fbc5e685a5d3d45aa1d0347fdb7c4d35> as C<fbc?>.
=item --query
type: string