4.2 Data Import Wizard vs. Data Loader

Data Import Wizard vs. Data Loader

Salesforce provides two primary administrative tools for importing, updating, and exporting relational data: the browser-based Data Import Wizard and the desktop/command-line Data Loader. Choosing the right tool requires evaluating specific platform boundaries and operational constraints.

---

Data Import Wizard: Architecture & Scope

The Data Import Wizard is an integrated, browser-based administrative interface accessible directly within Setup (Setup > Integrations > Data Import Wizard). It is engineered for business users and administrators executing small-to-medium ad-hoc data loads without requiring local client software installation.

Key Capabilities & Object Support

• Record Limit: Supports imports of up to 50,000 records in a single execution.
• Supported Standard Objects: Strictly limited to Accounts, Contacts, Leads, Solutions, and Campaign Members.
• Custom Objects: Supports data loading for supported custom objects.
• Unsupported Objects (Exam Trap): The Data Import Wizard CANNOT import or update Opportunities, Cases, Products, Assets, Users, Quotes, or Campaigns (as parent campaign records). If a requirement involves importing Opportunities or Cases, Data Loader is mandatory.

Built-In Deduplication Matching

A primary architectural advantage of the Data Import Wizard is native deduplication matching during the import process, which prevents duplicate record creation declaratively without requiring external ETL staging:

• Accounts & Contacts: Matches incoming rows against existing database records by Salesforce ID, Name & Site, Email, or designated External ID.
• Leads: Matches incoming leads by Salesforce ID, Email, or Name.

Administrators can configure the wizard to either update existing matching records, create new records, or ignore duplicates entirely.

Workflow and Automation Execution

The Data Import Wizard includes a declarative configuration checkbox: "Trigger workflow rules and processes for new and updated records?" When unchecked, workflow rules, process builder processes, and flows that fire on record creation or edit are bypassed, speeding up execution and preventing mass notification emails from being dispatched to customers during data loading.

---

Data Loader: Enterprise Scale & Operations

Data Loader is a client application installed locally on Windows or macOS, and it also functions as a Windows-only supported Command Line Interface (CLI). It is engineered for enterprise-scale bulk operations, complex schema migrations, and automated batch processing.

Key Capabilities & Scope

• Record Limit: Supports bulk operations of up to 5,000,000 records per run.
• Universal Object Support: Operates across supported standard objects (including Opportunities, Cases, Products, Users, Tasks, Events) and supported custom objects.
• Full Operation Suite: Supports seven distinct operational modes:
  1. Insert: Creates new records from an input CSV file.
  2. Update: Modifies existing records based on matching 15-character or 18-character Salesforce Record IDs.
  3. Upsert: Matches incoming rows against either a Salesforce Record ID or an External ID field; updates existing matches and inserts non-matching rows.
  4. Delete: Moves matching records into the Recycle Bin (soft delete, recoverable for 15 days).
  5. Hard Delete: Permanently purges records immediately, completely bypassing the Recycle Bin. Requires the "Bulk API Hard Delete" user permission and enabling Bulk API in Data Loader Settings.
  6. Export: Queries and extracts active (non-deleted) records matching a custom SOQL query into a CSV file.
  7. Export All: Queries and extracts active records plus soft-deleted records currently residing in the Recycle Bin (WHERE isDeleted = true).

Automated Batch Jobs via CLI

Data Loader includes a Command Line Interface (CLI) that can be fully automated using configuration files (process-conf.xml, config.properties). Organizations routinely script Data Loader via Windows Task Scheduler or Windows Task Scheduler jobs to execute automated nightly imports, ERP delta synchronizations, and compliance data extractions without manual user intervention.

---

Comprehensive Architectural Comparison: Data Import Wizard vs. Data Loader

---

The Upsert Operation & External IDs

The Upsert operation chooses Insert or Update for each incoming record; a whole multi-batch job is not one atomic transaction. It is the gold standard for synchronizing Salesforce with external systems (such as SAP, NetSuite, or legacy databases):

1. During configuration, the administrator designates a unique key: either the internal Salesforce Record ID or a custom field flagged as an External ID.
2. When Data Loader reads each row in the CSV file, it evaluates the External ID:
   • Match Found: Updates the existing Salesforce record with incoming CSV values.
   • No Match Found: Creates a new record in Salesforce populated with CSV values.
   • Multiple Matches Found: Logs an error for that specific row (DUPLICATE_EXTERNAL_ID) and skips it without failing the rest of the batch.

---

Data Formatting & Large-Scale Loading Best Practices

High-volume data loads require rigorous data preparation and performance tuning to prevent operational failure:

• Date and Date/Time Formatting: ISO 8601 is a clear, portable choice for dates and DateTime values: YYYY-MM-DD (e.g., 2026-10-18) or YYYY-MM-DDThh:mm:ss.sssZ for date/time fields. Other supported formats depend on Data Loader settings; test parsing and timezone handling before the production load.
• Multi-Select Picklist Formatting: Multiple values must be delimited by semicolons without extraneous spaces: ValueA;ValueB;ValueC.
• Resolving Record Locking (UNABLE_TO_LOCK_ROW): When inserting hundreds of thousands of child records (e.g., Contacts, Order Items, Transactions) linked to parent records, Bulk API processes multiple batches in parallel threads. If multiple batches attempt to lock the same parent record simultaneously to update system fields or recalculate Roll-Up Summary aggregates, lock contention occurs, throwing UNABLE_TO_LOCK_ROW errors.

Mitigating Lock Contention:

1. Sort the CSV File: Sort the input CSV file alphabetically or numerically by the Parent Record ID column. This reduces the likelihood of parallel batches competing for a parent; very large parent groups can still cross batch boundaries.
2. Enable Serial Mode: In Data Loader Settings, enable "Use Bulk API" and check "Enable Serial Mode". Serial mode forces batches to process sequentially one after another, eliminating multi-threaded race conditions at the expense of longer load times.
3. Controlled Automation Bypass: Prefer a reviewed, narrowly assigned bypass switch using a Hierarchical Custom Setting ($Setup.AutomationBypass__c.BypassAll__c = TRUE).