Can bulk loader load object array without specified schema?

Users
,
1

If I don’t specify a predicate as a list in schema file, and just describe it with multiple objects in nq file, can bulk loader automatically load it as a array? Eg:
entity.rdf:
<0x01> alias “alias1”
<0x01> alias “alias2”
<0x01> name “name”
<0x01> uid “myid”
schema.rdf:
uid: string @index(exact)
name: string @index(exact)

then run:
dgraph bulk -f ./entity.rdf -s ./schema.rdf --map_shards=2 --reduce_shards=1 --http localhost:8001 --zero=localhost:5080

It seems not working. Is there anyway to support it because I have thousands predicates and just want to specify severals in schema file instead enumerating all, the rest can be loaded as string or string list.

2

Can bulk loader load object array without specified schema?

Short answer, yes, but you shouldn’t.

You can add a “dummy” schema or a schema without several definitions and types. That’s okay. But not a good practice.

If your dataset is like this. Dgraph will recognize it as a List Type.

Remove these flags. It will run the defaults one (that’s not mandatory)

humm, okay. I gonna check it today. It should work.

3

If the object is uid/blanknode type, bulk loader can recognize it as a list of edges, otherwise it cannot.

4

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.