src-d · ajnavarro · Oct 31, 2018 · Oct 21, 2018 · Oct 21, 2018 · Oct 25, 2018
@@ -67,6 +67,7 @@ We support and actively test against certain third-party clients to ensure compa
 - `FLOOR(number)`: Return the largest integer value that is less than or equal to `number`.
 - `ROUND(number, decimals)`: Round the `number` to `decimals` decimal places.
 - `CONNECTION_ID()`: Return the current connection ID.
+- `SOUNDEX(str)`: Returns the soundex of a string.
 
 ## Example
 

@@ -87,6 +87,7 @@
 - ROUND
 - COALESCE
 - CONNECTION_ID
+- SOUNDEX
 
 ## Time functions
 - DAY

@@ -45,4 +45,5 @@ var Defaults = sql.Functions{
 	"coalesce":      sql.FunctionN(NewCoalesce),
 	"json_extract":  sql.FunctionN(NewJSONExtract),
 	"connection_id": sql.Function0(NewConnectionID),
+	"soundex":       sql.Function1(NewSoundex),
 }
diff --git a/sql/expression/function/soundex.go b/sql/expression/function/soundex.go
@@ -0,0 +1,102 @@
+package function
+
+import (
+	"fmt"
+	"strings"
+	"unicode"
+
+	"gopkg.in/src-d/go-mysql-server.v0/sql"
+	"gopkg.in/src-d/go-mysql-server.v0/sql/expression"
+)
+
+// Soundex is a function that returns the soundex of a string. Two strings that
+// sound almost the same should have identical soundex strings. A standard
+// soundex string is four characters long, but the SOUNDEX() function returns
+// an arbitrarily long string.
+type Soundex struct {
+	expression.UnaryExpression
+}
+
+// NewSoundex creates a new Soundex expression.
+func NewSoundex(e sql.Expression) sql.Expression {
+	return &Soundex{expression.UnaryExpression{Child: e}}
+}
+
+// Eval implements the Expression interface.
+func (s *Soundex) Eval(ctx *sql.Context, row sql.Row) (interface{}, error) {
+	v, err := s.Child.Eval(ctx, row)
+	if err != nil {
+		return nil, err
+	}
+
+	if v == nil {
+		return nil, nil
+	}
+
+	v, err = sql.Text.Convert(v)
+	if err != nil {
+		return nil, err
+	}
+
+	var b strings.Builder
+	var last rune
+	for _, c := range strings.ToUpper(v.(string)) {
+		if last == 0 && !unicode.IsLetter(c) {
+			continue
+		}
+		code := s.code(c)
+		if last == 0 {
+			b.WriteRune(c)
+			last = code
+			continue
+		}
+		if code == '0' || code == last {
+			continue
+		}
+		b.WriteRune(code)
+		last = code
+	}
+	if b.Len() == 0 {
+		return "", nil
+	}
+	for i := len([]rune(b.String())); i < 4; i++ {
+		b.WriteRune('0')
+	}
+	return b.String(), nil
+}
+
+func (s *Soundex) code(c rune) rune {
+	switch c {
+	case 'B', 'F', 'P', 'V':
+		return '1'
+	case 'C', 'G', 'J', 'K', 'Q', 'S', 'X', 'Z':
+		return '2'
+	case 'D', 'T':
+		return '3'
+	case 'L':
+		return '4'
+	case 'M', 'N':
+		return '5'
+	case 'R':
+		return '6'
+	}
+	return '0'
+}
+
+func (s *Soundex) String() string {
+	return fmt.Sprintf("SOUNDEX(%s)", s.Child)
+}
+
+// TransformUp implements the Expression interface.
+func (s *Soundex) TransformUp(f sql.TransformExprFunc) (sql.Expression, error) {
+	child, err := s.Child.TransformUp(f)
+	if err != nil {
+		return nil, err
+	}
+	return f(NewSoundex(child))
+}
+
+// Type implements the Expression interface.
+func (s *Soundex) Type() sql.Type {
+	return s.Child.Type()
+}
diff --git a/sql/expression/function/soundex_test.go b/sql/expression/function/soundex_test.go
@@ -0,0 +1,49 @@
+package function
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/require"
+	"gopkg.in/src-d/go-mysql-server.v0/sql"
+	"gopkg.in/src-d/go-mysql-server.v0/sql/expression"
+)
+
+func TestSoundex(t *testing.T) {
+	testCases := []struct {
+		name     string
+		rowType  sql.Type
+		row      sql.Row
+		expected interface{}
+	}{
+		{"text nil", sql.Text, sql.NewRow(nil), nil},
+		{"text empty", sql.Text, sql.NewRow(""), ""},
+		{"text ignored character", sql.Text, sql.NewRow("-"), ""},
+		{"text runes", sql.Text, sql.NewRow("日本語"), "日000"},
+		{"text Hello ok", sql.Text, sql.NewRow("Hello"), "H400"},
+		{"text Quadratically ok", sql.Text, sql.NewRow("Quadratically"), "Q36324"},
+		{"text Lee ok", sql.Text, sql.NewRow("Lee"), "L000"},
+		{"text McKnockitter ok", sql.Text, sql.NewRow("McKnockitter"), "M25236"},
+		{"text Honeyman ok", sql.Text, sql.NewRow("Honeyman"), "H500"},
+		{"text Munn ok", sql.Text, sql.NewRow("Munn"), "M000"},
+		{"text Poppett ok", sql.Text, sql.NewRow("Poppett"), "P300"},
+		{"text Peachman ok", sql.Text, sql.NewRow("Peachman"), "P250"},
+		{"text Cochrane ok", sql.Text, sql.NewRow("Cochrane"), "C650"},
+		{"text Chesley ok", sql.Text, sql.NewRow("Chesley"), "C400"},
+		{"text Tachenion ok", sql.Text, sql.NewRow("Tachenion"), "T250"},
+		{"text Wilcox ok", sql.Text, sql.NewRow("Wilcox"), "W420"},
+		{"binary ok", sql.Blob, sql.NewRow([]byte("Harvey")), "H610"},
+		{"other type", sql.Int32, sql.NewRow(int32(1)), ""},
+	}
+
+	for _, tt := range testCases {
+		f := NewSoundex(expression.NewGetField(0, tt.rowType, "", true))
+
+		t.Run(tt.name, func(t *testing.T) {
+			require.Equal(t, tt.expected, eval(t, f, tt.row))
+		})
+
+		req := require.New(t)
+		req.True(f.IsNullable())
+		req.Equal(tt.rowType, f.Type())
+	}
+}
-Original file line number
+Diff line change
@@ Expand Up / @@ -87,6 +87,7 @@ @@
     - ROUND
     - COALESCE
     - CONNECTION_ID
+    - SOUNDEX
     ## Time functions
     - DAY
@@ Expand Down @@