Connect NetSuite to Snowflake

NetSuite is the source of truth for revenue, cost, and inventory. Snowflake is where finance and analytics model it against sales, product, and marketing. NetSuite has no change-data-capture stream, so the usual pipeline is a nightly saved-search export, a fragile ODBC job, or a managed connector that is a black box at month-end.

• Transactions and entities land in Snowflake every few minutes on a lastModified watermark, and the connector's Record changed trigger polls a SuiteQL last-modified timestamp for the tables that need a shorter interval
• SuiteQL joins pull a transaction with its customer, subsidiary, and line items in one query, so the warehouse tables are already shaped for reporting
• Every load is an ordered run log with the query, the row count, and the merge, so a number that looks wrong in a board is traceable to a specific sync
• Concurrency governance is respected: the connector paces queries so a backfill does not trip NetSuite limits for the rest of the account

The load flow most finance-analytics teams turn on first.

NetSuite transactions into RAW.NETSUITE.TRANSACTION, every five minutes

Watermark-driven through SuiteQL, with a nightly full reconcile.

1. 1Trigger: a schedule every 5 minutes.
2. 2Query: Run SuiteQL for transactions where lastmodifieddate > :watermark, joining the entity, subsidiary, and currency, up to 1,000 rows per page.
3. 3Lines: a second SuiteQL query pulls the transaction lines for the same set, into a TRANSACTION_LINE table.
4. 4Stage and merge: for a batch above 5,000 rows, stage and COPY INTO; below that, merge directly. MERGE INTO RAW.NETSUITE.TRANSACTION t USING :batch s ON t.INTERNAL_ID = s.INTERNAL_ID.
5. 5Watermark: write the max lastmodifieddate seen to a control table.
6. 6Deletes: a nightly saved search of deleted records marks IS_DELETED = TRUE in the warehouse.
7. 7Nightly reconcile: a full saved-search export of the trailing 45 days is merged, catching any row a short-window watermark missed.

The starting map for the transaction to RAW.NETSUITE.TRANSACTION step.

• Transactions and lines. Sales orders, invoices, bills, journals, and their lines land as typed tables, joined to entity and subsidiary. (NetSuite to Snowflake)
• Entities and items. Customers, vendors, employees, and the item list sync as dimension tables for the fact tables to join. (NetSuite to Snowflake)
• Custom records. Any custom record type is picked up from the metadata catalog and loaded the same way as a standard record. (NetSuite to Snowflake)
• Saved-search results. An existing saved search runs on a schedule and its result set lands as a table, keeping the report logic in NetSuite. (NetSuite to Snowflake)

The same controls apply to this sync as to every other Neblex workload.

• Ordered run log. Every load records the SuiteQL statement, the row count, and the merge. Replay without re-querying NetSuite.
• Paced against governance. Queries are queued against the account concurrency limit, so a backfill runs slower rather than blocking other integrations.
• Scoped access. NetSuite uses a role with read access to the record types being loaded. Snowflake uses a role with grants on the target schema only.
• Environments. Build against a NetSuite sandbox and a Snowflake scratch schema, then promote the same flow to production.