When a query is > 10Kb we first limit size of all constant string parameters

to 30 charaters and then the query is truncated to 10Kb. This prevent
pgbadger to waste time/hang with very long queries when inserting bytea
for example. The 10Kb limit can be controled witgh the --maxlength
command line parameter.

diff --git a/pgbadger b/pgbadger
index be84e708783b022e36fe75e44d8d227c4f881787..57c950b411531fbd5ac4b332d68264b139c2104e 100755 (executable)
--- a/pgbadger
+++ b/pgbadger
@@ -276,7 +276,7 @@ my $ident                   = '';
 my $top                     = 0;
 my $sample                  = 3;
 my $extension               = '';
-my $maxlength               = 0;
+my $maxlength               = 10240;
 my $graph                   = 1;
 my $nograph                 = 0;
 my $debug                   = 0;
@@ -345,6 +345,7 @@ my $json_prettify           = 0;
 my $is_tsung_output         = 0;
 my $report_per_database     = 0;
 my $html_outdir             = '';
+my $param_size_limit        = 24;
 
 my $NUMPROGRESS = 10000;
 my @DIMENSIONS  = (800, 300);
@@ -4320,7 +4321,48 @@ sub anonymize_query
        return $orig_query;
 }
 
-# Format numbers with comma for better reading
+# Limit size of SQL queries by truncating parameters values
+sub limit_param_size
+{
+       my $orig_query = shift;
+
+       return if (!$orig_query);
+
+       # Variable to hold parameters
+       my @param_cache = ();
+
+       # Remove comments
+       $orig_query =~ s/\/\*(.*?)\*\///gs;
+
+       # Clean query
+       $orig_query =~ s/\\'//g;
+       $orig_query =~ s/('')+//g;
+
+       # Anonymize each values
+       my $i = 0;
+       while ($orig_query =~ s/([^\s]+[\s\(]*)'([^']*)'/$1\%PGB$i\%/s)
+       {
+               push(@param_cache, $2);
+               $i++;
+       }
+
+       # Limit size of paramaters
+       for ($i = 0; $i <= $#param_cache; $i++)
+       {
+               if (length($param_cache[$i]) > $param_size_limit)
+               {
+                       $param_cache[$i] = substr($param_cache[$i], 0, $param_size_limit) . '[...]';
+               }
+       }
+
+       # Restore parameters
+       $orig_query =~ s/\%PGB(\d+)\%/'$param_cache[$1]'/g;
+
+       return $orig_query;
+}
+
+
+# Format numbers with comma for betterparam_cache reading
 sub comma_numbers
 {
        return 0 if ($#_ < 0);
@@ -14554,6 +14596,12 @@ sub store_queries
                        return 1;
                }
 
+               # First limit query size at parameter length
+               if ($maxlength > 0 && (length($cur_info{$t_pid}{query}) > $maxlength))
+               {
+                       $cur_info{$t_pid}{query} = &limit_param_size($cur_info{$t_pid}{query});
+               }
+
                # Truncate the query if requested by the user
                $cur_info{$t_pid}{query} = substr($cur_info{$t_pid}{query}, 0, $maxlength) . '[...]'
                        if (($maxlength > 0) && (length($cur_info{$t_pid}{query}) > $maxlength));
@@ -14563,6 +14611,12 @@ sub store_queries
        # We only process stored object with query here
        if ($cur_info{$t_pid}{statement})
        {
+               # First limit statement size at parameter length
+               if ($maxlength > 0 && (length($cur_info{$t_pid}{statement}) > $maxlength))
+               {
+                       $cur_info{$t_pid}{statement} = &limit_param_size($cur_info{$t_pid}{statement});
+               }
+
                # Truncate the statement if requested by the user
                $cur_info{$t_pid}{statement} = substr($cur_info{$t_pid}{statement}, 0, $maxlength) . '[...]'
                        if (($maxlength > 0) && (length($cur_info{$t_pid}{statement}) > $maxlength));
@@ -14874,6 +14928,16 @@ sub store_temporary_and_lock_infos
                        $cur_temp_info{$t_pid}{query} = &anonymize_query($cur_temp_info{$t_pid}{query});
                }
 
+               # First limit query size at parameter length
+               if ($maxlength > 0 && (length($cur_temp_info{$t_pid}{query}) > $maxlength))
+               {
+                       $cur_temp_info{$t_pid}{query} = &limit_param_size($cur_temp_info{$t_pid}{query});
+               }
+
+               # Truncate the query if requested by the user
+               $cur_temp_info{$t_pid}{query} = substr($cur_temp_info{$t_pid}{query}, 0, $maxlength) . '[...]'
+                       if (($maxlength > 0) && (length($cur_temp_info{$t_pid}{query}) > $maxlength));
+
                # Normalize query
                my $normalized = &normalize_query($cur_temp_info{$t_pid}{query});
 
@@ -14925,6 +14989,16 @@ sub store_temporary_and_lock_infos
                        $cur_lock_info{$t_pid}{query} = &anonymize_query($cur_lock_info{$t_pid}{query});
                }
 
+               # First limit query size at parameter length
+               if ($maxlength > 0 && (length($cur_lock_info{$t_pid}{query}) > $maxlength))
+               {
+                       $cur_lock_info{$t_pid}{query} = &limit_param_size($cur_lock_info{$t_pid}{query});
+               }
+
+               # Truncate the query if requested by the user
+               $cur_lock_info{$t_pid}{query} = substr($cur_lock_info{$t_pid}{query}, 0, $maxlength) . '[...]'
+                       if (($maxlength > 0) && (length($cur_lock_info{$t_pid}{query}) > $maxlength));
+
                # Normalize query
                my $normalized = &normalize_query($cur_lock_info{$t_pid}{query});
 
@@ -14971,6 +15045,16 @@ sub store_temporary_and_lock_infos
                        $cur_cancel_info{$t_pid}{query} = &anonymize_query($cur_cancel_info{$t_pid}{query});
                }
 
+               # First limit query size at parameter length
+               if ($maxlength > 0 && (length($cur_cancel_info{$t_pid}{query}) > $maxlength))
+               {
+                       $cur_cancel_info{$t_pid}{query} = &limit_param_size($cur_cancel_info{$t_pid}{query});
+               }
+
+               # Truncate the query if requested by the user
+               $cur_cancel_info{$t_pid}{query} = substr($cur_cancel_info{$t_pid}{query}, 0, $maxlength) . '[...]'
+                       if (($maxlength > 0) && (length($cur_cancel_info{$t_pid}{query}) > $maxlength));
+
                # Normalize query
                my $normalized = &normalize_query($cur_cancel_info{$t_pid}{query});