Graph Formats

Graph input format

This page specifies proposed standard format for both normal graphs and property graphs.

• Adjacency List

  • this is file where each line is of the format, src sink1 sink2 ...

• Adjacency List with additional information for loading

  • this is file where each line of the format, src pid sink1 sink2 ...
  • Any extra information required with sink?

• Adjacency List with Subgraph information including along with partition information

  • this is file where each line of the format, src pid sgid sink1 sink2 ...

• JSON format for property graphs

  • first line of the file contains partition id
  • Json Format each line of file is of the format, this is one of the defaults in Giraph(Code re-usability?)
    [srcid, srcvalue, sinkid1,edgeid1,edgevalue1], [sinkid2,edgeid2,edgevalue2] ...
  • Adding partition information for loading
    [srcid,pid, srcvalue, sinkid1,edgeid1,edgevalue1], [sinkid2,edgeid2,edgevalue2] ...
  • Adding partition and subgraph information for loading
    [srcid,pid,sgid,srcvalue, sinkid1,edgeid1,edgevalue1], [sinkid2,edgeid2,edgevalue2] ...
  • Value format needed for property graphs
    key1:type1:value1$ key2:type2:value2$.....