GraphAcademy

Importing CSV data into Neo4j

Importing Data

Creating Nodes, Properties, and Relationships

Data Types, Lists, and Labels

Importing data considerations

Transactions

The Cypher queries you have written will run within a single transaction. As a result, the data is rolled back if a failure occurs; and the graph is unchanged.

Importing significant volumes of data in a single transaction can result in large write operations - this can cause performance issues and potential failure.

You can split a query into multiple transactions using the CALL clause with IN TRANSACTIONS.

cypher
CALL {
  // query
} IN TRANSACTIONS [OF X ROWS]

For example, the following query would create the Person nodes in individual transactions.

cypher
CALL {
    LOAD CSV WITH HEADERS
    FROM 'https://data.neo4j.com/importing-cypher/persons.csv'
    AS row
    MERGE (p:Person {tmdbId: toInteger(row.person_tmdbId)})
    SET
    p.imdbId = toInteger(row.person_imdbId),
    p.bornIn = row.bornIn,
    p.name = row.name,
    p.bio = row.bio,
    p.poster = row.poster,
    p.url = row.url,
    p.born = date(row.born),
    p.died = date(row.died);
} IN TRANSACTIONS

You can batch the transactions by specifying the number of rows to process in each transaction.

For example, modifying the query above to process 100 rows in each transaction:

cypher
} IN TRANSACTIONS OF 100 ROWS

Check Your Understanding

Batching transactions

Complete the following Cypher statement to run a query in batches of 1000 rows.

cypher
CALL {
  // query
/*select:} IN TRANSACTIONS OF 1000 ROWS*/

  • ❏ } BY TRANSACTIONS OF 1000

  • ❏ } BY TRANSACTIONS OF 1000 ROWS

  • ❏ } IN TRANSACTIONS OF 1000

  • ✓ } IN TRANSACTIONS OF 1000 ROWS

Hint

You need to specify the number of rows in each transaction.

Solution

The correct syntax is:

cypher
CALL {
  // query
} IN TRANSACTIONS OF 1000 ROWS

Summary

In this lesson, you learned how to split a query into multiple transactions using the CALL clause with IN TRANSACTIONS.

In the next lesson, you will learn the importance of splitting your import into multiple steps and how to avoid the Eager problem.