Exporting Job Execution Results to Files (2.2.1)

Function

This API is used to export the execution result (result) of an asynchronous job (jobId) to a file.

• The following algorithms are supported:
  • PageRank, PersonalRank, and Pixie
  • Louvain, Label Propagation, and Connected Component
  • K-Core
  • SSSP, Shortest Path (including Time Window Shortest Path), Shortest Path of Vertex Sets, All Shortest Paths, and n Paths
  • Triangle Count, Cluster Coefficient, Degree Correlation, and Closeness
  • Link Prediction
  • Betweenness, edge_betweenness, and od_betweenness
• The following queries are supported:
  • Cypher Queries (2.2.16)
  • Querying Vertices That Meet Filter Criteria (1.0.0)
  • Querying Edges That Meet Filter Criteria (1.0.0)

URI

Request Parameters

• paginate parameter descriptions
  1. When pagination is enabled, fileName in the request body indicates the directory name, and the directory is used to store pagination files. When pagination is disabled, fileName indicates the file name. Before the export, ensure that the path the file name points to is empty so that the existing data on OBS will not be overwritten during the export.
  2. If the value of numThread is greater than the number of vCPUs used by the GES graph instance, the parameter is set to the number of vCPUs.
  3. The rowCountPerFile value affects the number of actually used threads. That is, when the ratio of the result set size to rowCountPerFile is less than numThread, the ratio is used as the value of numThread.
  4. If the request is canceled by the user, the data uploaded to OBS will not be deleted. For details about the API for canceling jobs, see Canceling a Job (1.0.0).
• When the pagination function is enabled, the file name is named as follows:

  If enable in the paginate parameter is set to true, fileName indicates a directory. Files in the directory are named using the combination of the thread ID and file number, and are separated by a period (.). For example, for 3.2 million data records, the examples of the first and last file names in different configurations are as follows:

  rowCountPerFile	100,000	100,000	1 million	1 million	5 million
  numThread	2	48	2	5	2
  Threads Actually Used	2	32	2	4	1
  Files Generated by a Single Thread	16	1	2	1	1
  First File Name	00.000.txt	00.000.txt	00.00.txt	00.00.txt	00.00.txt
  Last File Name	01.015.txt	031.000.txt	01.01.txt	03.00.txt	00.00.txt

Response Parameters

Example Request

• Example request 2: Export the execution results of an asynchronous job to an OBS file. The export path is demo_movie/. The name of the exported file is louvain. After the export, the original job is deleted.

  POST /ges/v1.0/{project_id}/graphs/{graph_name}/jobs/{job_id}/action?action_id=export-result
  {
      "exportPath": "demo_movie/",
      "fileName": "louvain",
      "erase":  true,
      "obsParameters": {
          "accessKey": "xxxx",
          "secretKey": "xxxx"
      }
  }

• Example request 2: Export the execution result of an asynchronous job to an OBS file. The export path is demo_movie/. The name of the exported file is louvain. By default, pagination is disabled. When data is exported by page, the maximum number of rows in each file is 100,000.

  POST /ges/v1.0/{project_id}/graphs/{graph_name}/jobs/{job_id}/action?action_id=export-result
  {
      "exportPath": "demo_movie/",
      "fileName": "louvain",
      "paginate":{
      "enable":true,
      "numThread":2,
      "rowCountPerFile":100000,
      },
      "obsParameters": {
          "accessKey": "xxxx",
          "secretKey": "xxxx"
      }
  }

  

  Currently, Cypher statements can only be used to export common value types, such as attribute values, numbers, and strings, but not composite value types (such as lists and maps), vertices, or edges. Example:

  • The results of the following statements can be exported:

    match (n) return id(n) limit 10

    match (n) return n.age, n.occupation

    match (n)-[r]->(m) return n.Rating limit 10

    unwind [1,2,3] as p return p

  • The exported TXT file contains null values or blank lines because the results of the following statements contain objects or compound values:

    return [1,2,3], {a:1}

    match (n) return n limit 10

    match (n)-[r]->(m) return r limit 10

Example Response

Status code: 200

Example response for a successful request

Http Status Code: 200
{
    "jobId": "2aa37d87-fb5a-44c4-81a1-c3876247d1cf000003836",
    "jobType": 15
}

Status code: 400

Example response for a failed request

Http Status Code: 400
{
    "jobId": "2aa37d87-fb5a-44c4-81a1-c3876247d1cf000003836",
    "jobType": 15
}

Status Codes

Error Codes

See Error Codes.

Exporting the Result in TXT File

1. The following is an example of the algorithm execution result, for example, content of Louvain.txt:

   # modularity: 0.4269691347613425,
   #community_num: 4,
   #runtime: 0.003784,
   #data_total_size: 34
   #community:
   1,1
   2,1
   ...

2. The following is an example Cypher query result:
   • Example 1

     Query statement:

     match (n:user)-[r]->(m:movie) return id(n),n.Name, n.Occupation, n.Age,r.Score,m.ChineseTitle

     Result:

     #data_total_size:1209
     #data_return_size:1209
     #data_offset:0
     #records:
     Vivian, artist, 25-34, 5, Lethal Weapon
     Vivian, Artist, 25-34, 4, Raising Arizona
     Mercedes, K-12 student, Under 18, 3, Lethal Weapon
     Mercedes, K-12 student, Under 18, 3, The Rock
     ...

   • Example 2

     Query statement:

     match (n)-->(m) where id(n)='Vivian' return labels(m),count(*)

     Result:

     #data_total_size:2
     #data_return_size:2
     #data_offset:0
     #records:
     user,5
     movie,2