support default value while import

README.md

@@ -196,6 +196,7 @@ schema:
           - name: age
             type: int
             index: 2
+            default: 1
           - name: name
             type: string
             index: 1
@@ -223,6 +224,7 @@ Each tag contains the following two properties:
   * `name`: **Required**. The property name, must be the same with the tag property in Nebula Graph.
   * `type`: **Optional**. The property type, currently  `bool`, `int`, `float`, `double`, `string`, `time`, `timestamp`, `date`, `datetime`, `geography`, `geography(point)`, `geography(linestring)` and `geography(polygon)` are supported.
   * `index`: **Optional**. The column number in the CSV file.
+  * `default`: **Optional**. use this default value if the column is missing in CSV file
 
 > **NOTE**: The properties in the preceding `prop` parameter must be sorted in the **same** way as in the CSV data file.
 

README_zh-CN.md

@@ -170,6 +170,7 @@ schema:
           - name: age
             type: int
             index: 2
+            default: 18
           - name: name
             type: string
             index: 1
@@ -196,6 +197,7 @@ schema:
   - `name`：**必填**。属性名称，同 Nebula Graph 中创建的 TAG 的属性名称一致。
   - `type`：**必填**。属性类型，目前支持 `bool`、`int`、`float`、`double`、`timestamp`、`string`、`geography`、`geography(point)`、`geography(linestring)`和`geography(polygon)` 几种类型。
   - `index`：**可选**。在 CSV 文件中的列标。
+  - `default`: **可选**。如果在csv文件中该列缺失的时候，使用该默认值
 
 > **注意**：上述 `props` 中的属性描述**顺序**必须同数据文件中的对应数据排列顺序一致。
 

pkg/config/config.go

@@ -43,9 +43,10 @@ type NebulaClientSettings struct {
 }
 
 type Prop struct {
-	Name  *string `json:"name" yaml:"name"`
-	Type  *string `json:"type" yaml:"type"`
-	Index *int    `json:"index" yaml:"index"`
+	Name       *string `json:"name" yaml:"name"`
+	Type       *string `json:"type" yaml:"type"`
+	Index      *int    `json:"index" yaml:"index"`
+	DefaultVal *string `json:"default" yaml:"default"`
 }
 
 type VID struct {
@@ -815,6 +816,9 @@ func (p *Prop) FormatValue(record base.Record) (string, error) {
 		return "", fmt.Errorf("Prop index %d out range %d of record(%v)", *p.Index, len(record), record)
 	}
 	r := record[*p.Index]
+	if r == "" && *p.DefaultVal != "" {
+		r = *p.DefaultVal
+	}
 	if p.IsStringType() {
 		return fmt.Sprintf("%q", r), nil
 	}
@@ -830,7 +834,7 @@ func (p *Prop) FormatValue(record base.Record) (string, error) {
 }
 
 func (p *Prop) String(prefix string) string {
-	return fmt.Sprintf("%s.%s:%s", prefix, *p.Name, *p.Type)
+	return fmt.Sprintf("%s.%s:%s:%s", prefix, *p.Name, *p.Type, *p.DefaultVal)
 }
 
 func (p *Prop) validateAndReset(prefix string, val int) error {
@@ -845,6 +849,11 @@ func (p *Prop) validateAndReset(prefix string, val int) error {
 			return fmt.Errorf("Invalid prop index: %d, name: %s, type: %s", *p.Index, *p.Name, *p.Type)
 		}
 	}
+
+	if p.DefaultVal == nil {
+		v := ""
+		p.DefaultVal = &v
+	}
 	return nil
 }
 

pkg/reader/batchmgr.go

@@ -107,31 +107,33 @@ func (bm *BatchMgr) InitSchema(header base.Record, runnerLogger *logger.RunnerLo
 }
 
 func (bm *BatchMgr) addVertexTags(r string, i int) {
-	columnName, columnType := bm.parseProperty(r)
+	columnName, columnType, defaultVal := bm.parseProperty(r)
 	tagName, prop := bm.parseTag(columnName)
 	if tagName == "" {
 		return
 	}
 	tag := bm.getOrCreateVertexTagByName(tagName)
 	p := config.Prop{
-		Name:  &prop,
-		Type:  &columnType,
-		Index: &i,
+		Name:       &prop,
+		Type:       &columnType,
+		Index:      &i,
+		DefaultVal: &defaultVal,
 	}
 	tag.Props = append(tag.Props, &p)
 }
 
 func (bm *BatchMgr) addEdgeProps(r string, i int) {
-	columnName, columnType := bm.parseProperty(r)
+	columnName, columnType, defaultVal := bm.parseProperty(r)
 	res := strings.SplitN(columnName, ".", 2)
 	prop := res[0]
 	if len(res) > 1 {
 		prop = res[1]
 	}
 	p := config.Prop{
-		Name:  &prop,
-		Type:  &columnType,
-		Index: &i,
+		Name:       &prop,
+		Type:       &columnType,
+		Index:      &i,
+		DefaultVal: &defaultVal,
 	}
 	bm.Schema.Edge.Props = append(bm.Schema.Edge.Props, &p)
 }
@@ -191,14 +193,25 @@ func (bm *BatchMgr) parseTag(s string) (tag, field string) {
 	return res[0], res[1]
 }
 
-func (bm *BatchMgr) parseProperty(r string) (columnName, columnType string) {
-	res := strings.SplitN(r, ":", 2)
+func (bm *BatchMgr) parseProperty(r string) (columnName, columnType, defaultVal string) {
+	res := strings.SplitN(r, ":", 3)
 
-	if len(res) == 1 || res[1] == "" || !base.IsValidType(res[1]) {
-		return res[0], "string"
+	if len(res) == 1 {
+		columnType = "string"
 	} else {
-		return res[0], res[1]
+		columnType = res[1]
 	}
+	if !base.IsValidType(columnType) {
+		columnType = "string"
+	}
+
+	if len(res) > 2 {
+		defaultVal = res[2]
+	} else {
+		defaultVal = ""
+	}
+
+	return res[0], columnType, defaultVal
 }
 
 func (bm *BatchMgr) Add(data base.Data, runnerLogger *logger.RunnerLogger) error {