Preserve UIDs in bulk load by default and add --new_uids option to override.

Author: Javier Alvarado

dgraph/cmd/bulk/loader.go

@@ -58,6 +58,7 @@ type options struct {
 	HttpAddr         string
 	IgnoreErrors     bool
 	CustomTokenizers string
+	NewUids          bool
 
 	MapShards    int
 	ReduceShards int

dgraph/cmd/bulk/mapper.go

@@ -26,6 +26,7 @@ import (
 	"os"
 	"path/filepath"
 	"sort"
+	"strconv"
 	"strings"
 	"sync"
 	"sync/atomic"
@@ -185,11 +186,11 @@ func (m *mapper) addMapEntry(key []byte, p *pb.Posting, shard int) {
 }
 
 func (m *mapper) processNQuad(nq gql.NQuad) {
-	sid := m.lookupUid(nq.GetSubject())
+	sid := m.uid(nq.GetSubject())
 	var oid uint64
 	var de *pb.DirectedEdge
 	if nq.GetObjectValue() == nil {
-		oid = m.lookupUid(nq.GetObjectId())
+		oid = m.uid(nq.GetObjectId())
 		de = nq.CreateUidEdge(sid, oid)
 	} else {
 		var err error
@@ -216,6 +217,17 @@ func (m *mapper) processNQuad(nq gql.NQuad) {
 	}
 }
 
+func (m *mapper) uid(xid string) uint64 {
+	if !m.opt.NewUids {
+		if uid, err := strconv.ParseUint(xid, 0, 64); err == nil {
+			m.xids.BumpTo(uid)
+			return uid
+		}
+	}
+
+	return m.lookupUid(xid)
+}
+
 func (m *mapper) lookupUid(xid string) uint64 {
 	uid := m.xids.AssignUid(xid)
 	if !m.opt.StoreXids {

dgraph/cmd/bulk/run.go

@@ -90,6 +90,8 @@ func init() {
 			"more parallelism, but increases memory usage.")
 	flag.String("custom_tokenizers", "",
 		"Comma separated list of tokenizer plugins")
+	flag.Bool("new_uids", false,
+		"Ignore UIDs in load files and assign new ones.")
 }
 
 func run() {
@@ -113,6 +115,7 @@ func run() {
 		MapShards:        Bulk.Conf.GetInt("map_shards"),
 		ReduceShards:     Bulk.Conf.GetInt("reduce_shards"),
 		CustomTokenizers: Bulk.Conf.GetString("custom_tokenizers"),
+		NewUids:          Bulk.Conf.GetBool("new_uids"),
 	}
 
 	x.PrintVersion()