Merge pull request #162 from Shopify/nested_stats2

Recursive/aggregate bucket stats

3 files changed, 147 insertions, 0 deletions

diff --git a/cmd/bolt/main.go b/cmd/bolt/main.go
index 82b5e5c..66c33d2 100644
--- a/cmd/bolt/main.go
+++ b/cmd/bolt/main.go
@@ -92,6 +92,14 @@ func NewApp() *cli.App {
 			},
 		},
 		{
+			Name:  "stats",
+			Usage: "Aggregate statistics for all buckets matching specified prefix",
+			Action: func(c *cli.Context) {
+				path, name := c.Args().Get(0), c.Args().Get(1)
+				Stats(path, name)
+			},
+		},
+		{
 			Name:  "bench",
 			Usage: "Performs a synthetic benchmark",
 			Flags: []cli.Flag{
diff --git a/cmd/bolt/stats.go b/cmd/bolt/stats.go
new file mode 100644
index 0000000..54a0f44
--- /dev/null
+++ b/cmd/bolt/stats.go
@@ -0,0 +1,77 @@
+package main
+
+import (
+	"bytes"
+	"os"
+
+	"github.com/boltdb/bolt"
+)
+
+// Collect stats for all top level buckets matching the prefix.
+func Stats(path, prefix string) {
+	if _, err := os.Stat(path); os.IsNotExist(err) {
+		fatal(err)
+		return
+	}
+
+	db, err := bolt.Open(path, 0600)
+	if err != nil {
+		fatal(err)
+		return
+	}
+	defer db.Close()
+
+	err = db.View(func(tx *bolt.Tx) error {
+		var s bolt.BucketStats
+		var count int
+		var prefix = []byte(prefix)
+		tx.ForEach(func(name []byte, b *bolt.Bucket) error {
+			if bytes.HasPrefix(name, prefix) {
+				s.Add(b.Stats())
+				count += 1
+			}
+			return nil
+		})
+		printf("Aggregate statistics for %d buckets\n\n", count)
+
+		println("Page count statistics")
+		printf("\tNumber of logical branch pages: %d\n", s.BranchPageN)
+		printf("\tNumber of physical branch overflow pages: %d\n", s.BranchOverflowN)
+		printf("\tNumber of logical leaf pages: %d\n", s.LeafPageN)
+		printf("\tNumber of physical leaf overflow pages: %d\n", s.LeafOverflowN)
+
+		println("Tree statistics")
+		printf("\tNumber of keys/value pairs: %d\n", s.KeyN)
+		printf("\tNumber of levels in B+tree: %d\n", s.Depth)
+
+		println("Page size utilization")
+		printf("\tBytes allocated for physical branch pages: %d\n", s.BranchAlloc)
+		var percentage int
+		if s.BranchAlloc != 0 {
+			percentage = int(float32(s.BranchInuse) * 100.0 / float32(s.BranchAlloc))
+		}
+		printf("\tBytes actually used for branch data: %d (%d%%)\n", s.BranchInuse, percentage)
+		printf("\tBytes allocated for physical leaf pages: %d\n", s.LeafAlloc)
+		percentage = 0
+		if s.LeafAlloc != 0 {
+			percentage = int(float32(s.LeafInuse) * 100.0 / float32(s.LeafAlloc))
+		}
+		printf("\tBytes actually used for leaf data: %d (%d%%)\n", s.LeafInuse, percentage)
+
+		println("Bucket statistics")
+		printf("\tTotal number of buckets: %d\n", s.BucketN)
+		percentage = int(float32(s.InlineBucketN) * 100.0 / float32(s.BucketN))
+		printf("\tTotal number on inlined buckets: %d (%d%%)\n", s.InlineBucketN, percentage)
+		percentage = 0
+		if s.LeafInuse != 0 {
+			percentage = int(float32(s.InlineBucketInuse) * 100.0 / float32(s.LeafInuse))
+		}
+		printf("\tBytes used for inlined buckets: %d (%d%%)\n", s.InlineBucketInuse, percentage)
+
+		return nil
+	})
+	if err != nil {
+		fatal(err)
+		return
+	}
+}
diff --git a/cmd/bolt/stats_test.go b/cmd/bolt/stats_test.go
new file mode 100644
index 0000000..2ad5d51
--- /dev/null
+++ b/cmd/bolt/stats_test.go
@@ -0,0 +1,62 @@
+package main_test
+
+import (
+	"os"
+	"strconv"
+	"testing"
+
+	"github.com/boltdb/bolt"
+	. "github.com/boltdb/bolt/cmd/bolt"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestStats(t *testing.T) {
+	if os.Getpagesize() != 4096 {
+		t.Skip()
+	}
+	SetTestMode(true)
+	open(func(db *bolt.DB, path string) {
+		db.Update(func(tx *bolt.Tx) error {
+			b, err := tx.CreateBucket([]byte("foo"))
+			if err != nil {
+				return err
+			}
+			for i := 0; i < 10; i++ {
+				b.Put([]byte(strconv.Itoa(i)), []byte(strconv.Itoa(i)))
+			}
+			b, err = tx.CreateBucket([]byte("bar"))
+			if err != nil {
+				return err
+			}
+			for i := 0; i < 100; i++ {
+				b.Put([]byte(strconv.Itoa(i)), []byte(strconv.Itoa(i)))
+			}
+			b, err = tx.CreateBucket([]byte("baz"))
+			if err != nil {
+				return err
+			}
+			b.Put([]byte("key"), []byte("value"))
+			return nil
+		})
+		db.Close()
+		output := run("stats", path, "b")
+		assert.Equal(t, "Aggregate statistics for 2 buckets\n\n"+
+			"Page count statistics\n"+
+			"\tNumber of logical branch pages: 0\n"+
+			"\tNumber of physical branch overflow pages: 0\n"+
+			"\tNumber of logical leaf pages: 1\n"+
+			"\tNumber of physical leaf overflow pages: 0\n"+
+			"Tree statistics\n"+
+			"\tNumber of keys/value pairs: 101\n"+
+			"\tNumber of levels in B+tree: 1\n"+
+			"Page size utilization\n"+
+			"\tBytes allocated for physical branch pages: 0\n"+
+			"\tBytes actually used for branch data: 0 (0%)\n"+
+			"\tBytes allocated for physical leaf pages: 4096\n"+
+			"\tBytes actually used for leaf data: 1996 (48%)\n"+
+			"Bucket statistics\n"+
+			"\tTotal number of buckets: 2\n"+
+			"\tTotal number on inlined buckets: 1 (50%)\n"+
+			"\tBytes used for inlined buckets: 40 (2%)", output)
+	})
+}