Open Hyundong-Seo opened 1 year ago
By definition, csv means a comma separated values file. I don't think it is wise to add in another delimiter.
If my understanding of the load is correct, all of the information needed to create a graph component (vertex or edge) is contained within the single line for that component. This precludes having it just in the header.
If I understand what you are suggesting, what would happen if there were many labels in that csv file?
commentid:comment3,personid:person
1236950581249,10995116284808
1236950581250,4139
2061584302085,4139
commentid:comment2,personid:person
2061584302086,4139
2061584302087,4139
commentid:comment2,personid:person2
2061584302088,10995116284808
2061584302089,32985348838375
commentid:comment1,personid:person3
2061584302090,10995116284808
2061584302091,6597069777240
This would make creating and parsing that file a real problem, without any real benefit.
Does this help?
When I try to load a csv file for edge, I should write a format like this for only AGE. start_id,start_vertex_type,end_id,end_vertex_type,properties
But I think that it wastes about data and it isn't reasonable, because i should modify all files that were already written It seems that they have to exist for only AGE and it is not suitable about reusing a data file.
I suggest a method that it changes just header not the other rows like this. And we write a label name with other delimiter like :(the colon). [present]
[suggestion]