From 6befaeecc9554be09177e699544d9573e699f1c8 Mon Sep 17 00:00:00 2001
From: Simon J Mudd <sjmudd@pobox.com>
Date: Sat, 22 Nov 2014 00:20:21 +0100
Subject: [PATCH] Add regexp caching to reduce load

---
 .../file_summary_by_instance/cache.go              | 36 ++++++++++++++++++++++
 .../file_summary_by_instance.go                    | 14 ++++++++-
 .../file_summary_by_instance_row.go                | 32 ++++++++++---------
 3 files changed, 67 insertions(+), 15 deletions(-)
 create mode 100644 performance_schema/file_summary_by_instance/cache.go

diff --git a/performance_schema/file_summary_by_instance/cache.go b/performance_schema/file_summary_by_instance/cache.go
new file mode 100644
index 0000000..6b2b147
--- /dev/null
+++ b/performance_schema/file_summary_by_instance/cache.go
@@ -0,0 +1,36 @@
+package file_summary_by_instance
+
+import (
+	"errors"
+
+	//	"github.com/sjmudd/pstop/lib"
+)
+
+// provide a mapping from filename to table.schema etc
+
+var (
+	mapped_name    map[string]string
+	total, matched int
+)
+
+func init() {
+	// setup on startup
+	mapped_name = make(map[string]string)
+}
+
+func get_from_cache(key string) (result string, err error) {
+	total++
+	if result, ok := mapped_name[key]; ok {
+		matched++
+		//		lib.Logger.Println("matched/total:", matched, total)
+		return result, nil
+	} else {
+		//		lib.Logger.Println("matched/total:", matched, total)
+		return "", errors.New("Not found")
+	}
+}
+
+func save_to_cache(key, value string) string {
+	mapped_name[key] = value
+	return value
+}
diff --git a/performance_schema/file_summary_by_instance/file_summary_by_instance.go b/performance_schema/file_summary_by_instance/file_summary_by_instance.go
index 10ef5f4..737caa8 100644
--- a/performance_schema/file_summary_by_instance/file_summary_by_instance.go
+++ b/performance_schema/file_summary_by_instance/file_summary_by_instance.go
@@ -6,6 +6,7 @@ package file_summary_by_instance
 
 import (
 	"database/sql"
+	"fmt"
 	"time"
 
 	"github.com/sjmudd/pstop/lib"
@@ -139,7 +140,8 @@ func (t File_summary_by_instance) EmptyRowContent() string {
 }
 
 func (t File_summary_by_instance) Description() string {
-	return "File I/O by filename (file_summary_by_instance)"
+	count := t.count_rows()
+	return fmt.Sprintf("File I/O by filename (file_summary_by_instance) %4d row(s)    ", count)
 }
 
 // create a new structure and include various variable values:
@@ -152,3 +154,13 @@ func NewFileSummaryByInstance(global_variables map[string]string) *File_summary_
 
 	return n
 }
+
+func (t File_summary_by_instance) count_rows() int {
+	var count int
+	for row := range t.results {
+		if t.results[row].SUM_TIMER_WAIT > 0 {
+			count++
+		}
+	}
+	return count
+}
diff --git a/performance_schema/file_summary_by_instance/file_summary_by_instance_row.go b/performance_schema/file_summary_by_instance/file_summary_by_instance_row.go
index 20bd606..6c10ef4 100644
--- a/performance_schema/file_summary_by_instance/file_summary_by_instance_row.go
+++ b/performance_schema/file_summary_by_instance/file_summary_by_instance_row.go
@@ -209,6 +209,10 @@ func (t file_summary_by_instance_row) simple_name(global_variables map[string]st
 
 	path := t.FILE_NAME
 
+	if cached_result, err := get_from_cache(path); err == nil {
+		return cached_result
+	}
+
 	// FIXME and make this work.
 	//	re4 := regexp.MustCompile(re_encoded)
 	//	if m4 := re4.FindStringSubmatch(path); m4 != nil {
@@ -222,33 +226,33 @@ func (t file_summary_by_instance_row) simple_name(global_variables map[string]st
 	if m1 := re_table_file.FindStringSubmatch(path); m1 != nil {
 		// we may match temporary tables so check for them
 		if m2 := re_temp_table.FindStringSubmatch(m1[2]); m2 != nil {
-			return "<temp_table>"
+			return save_to_cache(path, "<temp_table>")
 		}
 
 		// we may match partitioned tables so check for them
 		if m3 := re_part_table.FindStringSubmatch(m1[2]); m3 != nil {
-			return m1[1] + "." + m3[1] // <schema>.<table> (less partition info)
+			return save_to_cache(path, m1[1]+"."+m3[1]) // <schema>.<table> (less partition info)
 		}
 
-		return m1[1] + "." + m1[2] // <schema>.<table>
+		return save_to_cache(path, m1[1]+"."+m1[2]) // <schema>.<table>
 	}
 	if re_ibdata.MatchString(path) == true {
-		return "<ibdata>"
+		return save_to_cache(path, "<ibdata>")
 	}
 	if re_redo_log.MatchString(path) == true {
-		return "<redo_log>"
+		return save_to_cache(path, "<redo_log>")
 	}
 	if re_binlog.MatchString(path) == true {
-		return "<binlog>"
+		return save_to_cache(path, "<binlog>")
 	}
 	if re_db_opt.MatchString(path) == true {
-		return "<db_opt>"
+		return save_to_cache(path, "<db_opt>")
 	}
 	if re_slowlog.MatchString(path) == true {
-		return "<slow_log>"
+		return save_to_cache(path, "<slow_log>")
 	}
 	if re_auto_cnf.MatchString(path) == true {
-		return "<auto_cnf>"
+		return save_to_cache(path, "<auto_cnf>")
 	}
 	// relay logs are a bit complicated. If a full path then easy to
 	// identify,but if a relative path we may need to add $datadir,
@@ -261,19 +265,19 @@ func (t file_summary_by_instance_row) simple_name(global_variables map[string]st
 		}
 		re_relay_log := relay_log + `\.(\d{6}|index)$`
 		if regexp.MustCompile(re_relay_log).MatchString(path) == true {
-			return "<relay_log>"
+			return save_to_cache(path, "<relay_log>")
 		}
 	}
 	if re_pid_file.MatchString(path) == true {
-		return "<pid_file>"
+		return save_to_cache(path, "<pid_file>")
 	}
 	if re_error_msg.MatchString(path) == true {
-		return "<errmsg>"
+		return save_to_cache(path, "<errmsg>")
 	}
 	if re_charset.MatchString(path) == true {
-		return "<charset>"
+		return save_to_cache(path, "<charset>")
 	}
-	return path
+	return save_to_cache(path, path)
 }
 
 // Convert the imported "table" to a merged one with merged data.
-- 
2.7.4