Mac High Sierra环境下Neo4j的Load CSV导入命令语法咨询

Hey there! Let's walk through exactly how to structure the LOAD CSV command for your specific sales data CSV on Neo4j (running on Mac OS High Sierra). I'll break this down from the basic syntax to a fully customized example that matches your CSV fields.

Core LOAD CSV Structure #

The foundational syntax for LOAD CSV (perfect for your header-row CSV) looks like this:

LOAD CSV WITH HEADERS FROM 'file:///path/to/your/csv/file.csv' AS row
// Your logic here (create nodes, set properties, link relationships, etc.)

Key Notes for Mac OS: #

• Neo4j has a default import directory where it's safest to place your CSV file (avoids permission headaches). For Neo4j Desktop users, find it by right-clicking your database > "Open Folder" > "import".
• Once your CSV is in the import folder, you can reference it simply as file:///your_filename.csv—no need for the full system path.

Customized Command for Your CSV Schema #

Your CSV mixes numeric, string, and timestamp fields, so we'll cast values to Neo4j's native data types (since CSV reads all values as strings by default). Here's a ready-to-use example that creates a SalesRecord node for each row:

// Use PERIODIC COMMIT for large datasets to avoid memory overload
USING PERIODIC COMMIT 1000
LOAD CSV WITH HEADERS FROM 'file:///sales_data.csv' AS row
CREATE (s:SalesRecord {
    // Numeric fields: convert from string to integer
    salesVolumes: toInteger(row.SALES_VOLUMES),
    netSales: toInteger(row.NET_SALES),
    cm1: toInteger(row.CM1),
    // String fields: use directly
    productMainGroup: row.PRODUCT_MAIN_GROUP,
    region: row.REGION,
    sbu: row.SBU,
    salesType: row.SALES_TYPE,
    dataSource: row.DATA_SOURCE,
    period: row.PERIOD,
    // Period-related numeric fields
    periodYear: toInteger(row.PERIOD_YEAR),
    periodHalfYear: toInteger(row.PERIOD_HALF_YEAR),
    periodQuarter: toInteger(row.PERIOD_QUARTER),
    periodMonth: toInteger(row.PERIOD_MONTH),
    // Additional string fields
    currency: row.CURRENCY,
    userId: row.USERID,
    // Timestamp field: convert to Neo4j's datetime type
    entryTs: datetime(row.ENTRY_TS)
})

Optional Adjustments #

• Handling Null Values: If some fields might be empty, use coalesce() to set a default. For example:

  salesVolumes: coalesce(toInteger(row.SALES_VOLUMES), 0)
  

• Creating Relationships: If you want to link SalesRecord nodes to existing nodes (e.g., a Region node with name: row.REGION), modify the query to use MATCH + CREATE:

  USING PERIODIC COMMIT 1000
  LOAD CSV WITH HEADERS FROM 'file:///sales_data.csv' AS row
  MATCH (r:Region {name: row.REGION})
  CREATE (s:SalesRecord { /* ... properties ... */ })
  CREATE (s)-[:RECORDED_IN]->(r)
  

• Performance Tweaks: For very large CSVs, bump the PERIODIC COMMIT value (e.g., 5000) to speed up imports—just keep it aligned with your system's memory capacity.

内容的提问来源于stack exchange，提问作者David