Skip to content

Import data

After CSV files of data and a schema are created, you can use the Import page to batch import vertex and edge data into Nebula Graph for graph exploration and data analysis.

Prerequisites

To batch import data, do a check of these:

  • Studio is connected to Nebula Graph.
  • A schema is created.
  • CSV files meet the demands of the Schema.
  • Your account has privilege of GOD, ADMIN, DBA, or USER.

Procedure

To batch import data, follow these steps:

  1. In the toolbar, click the Import tab.

  2. On the Select Space page, choose a graph space name. In this example, basketballplayer is used. And then click the Next button.

  3. On the Upload Files page, click the Upload Files button and then choose CSV files. In this example, edge_serve.csv, edge_follow.csv, vertex_player.csv, and vertex_team.csv are chosen.

    Note

    You can choose multiple CSV files at the same time. The CSV file used in this article can be downloaded in the Design a schema.

  4. On the Select Files page, do a check of the file size and click Preview or Delete in the Operations column to make sure that all source data is correct. And then click the Next button.

  5. On the Map Vertices page, click the + Bind Datasource button, and in the dialog box, choose a CSV file. In this example, vertex_player.csv or vertex_team.csv is chosen.

  6. In the DataSource X tab, click the + Tag button.

  7. In the vertexId section, do these operations:

    a. In the CSV Index column, click Mapping.
    Click "Mapping** in the CSV Index column

    b. In the dialog box, choose a column from the CSV file. In this example, the only one cloumn of vertex_player.csv is chosen to generate VIDs representing players and the playerID column of vertex_player.csv is chosen to generate VIDs representing players.

    !!! Note

      In the same graph space, the VID is always unique and cannot be repeated. For VID information, see [VID](../../1.introduction/3.vid.md) "Click to enter the Nebula Graph Manual".
    
  8. In the TAG 1 section, do these operations:
    a. In the TAG drop-down list, choose a tag name. In this example, player is used for the vertex_player.csv file, and team is used for the vertex_team.csv file.

    b. In the property list, click Mapping to choose a data column from the CSV file as the value of a property. In this example, for the player tag, choose Column 1 for the age property and set its type to int. And choose Column 2 for the name property and set its type to string.

    Data source for the course vertices

  9. (Optional) If necessary, repeat Step 5 through Step 8 for more tags.

  10. When the configuration is done, click the Next button.
    When Config validation was successful prompts, data mapping for the vertices is successful.

  11. On the Map Edges page, click the + Bind Datasource button, and in the dialog box, choose a CSV file. In this example, the edge_follow.csv file is chosen.

  12. In the Type drop-down list, choose an edge type name. In this example, follow is chosen.

  13. In the property list, click Mapping to choose a column from the edge_follow.csv file as values of a property for the edges. srcId and dstId are the VIDs of the source vertex and destination vertex of an edge. In this example, srcId must be set to the VIDs of the player and dstId must be set to the VIDs of another player. Rank is optional.

    Data source for the action edges

  14. When the configuration is done, click the Next button.

  15. On the Import page, click the Start Import button. On the log window, you can see the import progress. The consumed time depends on the data volume. During data import, you can click the Stop Import button to stop data import. When the log window shows information as follows, the data import is done. The log window shows the time, imported rows, and operated lines

Next to do

When the data are imported to v2.6.0, you can query graph data.


Last update: September 6, 2021
Back to top