Salesforce Migration Validator – Data Reconciliation for QA
Did my Salesforce data migration land correctly?
Migration tools (Data Loader, Jitterbit, Gearset…) move data — they don't independently
prove it arrived right. This extension reconciles your source export (CSV) against the
target org, field by field, and reports what's missing, extra, truncated, mismatched, or
has broken lookups.
100% local · read-only · no backend · no telemetry. Your migration data (often customer
PII) never leaves your machine — the only network call is to your own org via the Salesforce
CLI you already use.
Built by a Salesforce QA lead for the exact question every migration raises: "has everything
migrated correctly?"
What it checks
- Counts: records that are missing (in the source but not in the org) and extra records.
- Field by field with normalisation (trim, upper/lower case, dates, numbers) → no false positives.
- Truncation and type mismatches.
- Relationships: lookups / master-detail resolve to an existing parent.
- Exportable report in Markdown / HTML / CSV to attach to the ticket. English and Spanish.
Checks in detail
| Group |
Checks |
| Counts |
Source rows vs org records · missing in the org · extra in the org · duplicate keys (source and org) · rows without a key |
| Field values |
Field-by-field comparison of every matched record, with type-aware normalisation so formatting differences don't show up as false positives |
| Relationships |
Lookups / master-detail resolve: lookup left empty, lookup pointing to another parent, parent record never migrated |
| Format & types |
Truncated values (org holds only the beginning), values longer than the field, values that can't be read as the field type (bad dates, numbers, booleans), values that aren't picklist values |
Normalisation (no false positives)
| Source |
Salesforce |
Equal? |
Ana |
Ana |
✅ trim |
ANA@EXAMPLE.COM |
ana@example.com |
✅ ignore case |
02/11/2026 (DD/MM/YYYY) |
2026-11-02 |
✅ date format |
2026-11-02 11:30 +01:00 |
2026-11-02T10:30:00.000Z |
✅ time zones |
1.234,50 € |
1234.5 |
✅ decimal separator, currency, field scale |
Y / sí / 1 |
true |
✅ checkbox |
G (map G=Gold) |
Gold |
✅ picklist value map |
b;a |
a;b |
✅ multi-select order |
001A0000006Vm9r |
001A0000006Vm9rIAC |
✅ 15/18-char Ids |
Privacy
No cloud. No account. No data upload. The extension runs entirely on your machine and only talks
to the org you're already authenticated against with the Salesforce CLI.
The mapping editor
Pick the org, the object and the source file. The editor suggests a mapping from column names and
field labels, picks an external ID as the matching key, and guesses date formats and decimal
separators from your data. Adjust it in one screen:
- Matching key: one column (ideally a legacy / external ID) or a composite key. Lookups can
be part of the key (e.g.
Account.Legacy_Id__c).
- Fields to compare: source column → Salesforce field, or a parent field through a lookup
(
Account.Legacy_Id__c, Owner.Username).
- Simple transformations per field: trim, upper/lower case, ignore case, date format, decimal
separator, value map (picklist codes, Y/N…).
- Org filter (optional SOQL
WHERE) to scope the records of this migration wave, e.g.
Migration_Batch__c = 'W1'.
Volume: REST or Bulk API 2.0
Small objects are read with SOQL over REST. When the object holds more than bulkThreshold
records (10,000 by default) the extension uses a Bulk API 2.0 query job (one job, results
streamed in pages), after checking your daily limits. Verified with 12,000 records on a real
scratch org: ~8 seconds, 12 API requests.
The report
- Verdict, tiles (source rows, org records, % matched, % fully equal, missing, extra), discrepancies
by check and by field, and the detail of each discrepancy with the source and org values.
- Export to Markdown, HTML or CSV to attach to the migration ticket. The CSV has every
discrepancy (one per line) and is protected against spreadsheet formula injection.
- English and Spanish.
Getting started
- Install the Salesforce CLI and sign in:
sf org login web --alias target-org
- In VS Code run Migration Validator: Validate a migration (CSV vs org) — from the Command
Palette, the Migration Validator icon in the Activity Bar, or right-click a
.csv file.
- Choose the org, the object and the source file. Check the mapping and the key, then Run
validation.
- Read the report; export it with the buttons at the top. Fix, re-load, and Re-run.
Supported source files: .csv, .tsv, .txt (delimiter detected: , ; tab |), and .json
(an array of objects, or { "records": [...] }). UTF-8 or Windows-1252. Excel: save as CSV UTF-8.
Security details
- Read-only. The extension only sends GET requests, plus the POST that creates a Bulk API query
job. It never creates, updates or deletes records.
- Your data stays here. Source files and org records are compared in memory on your computer.
Nothing is sent to any server of ours (there are none). Reports reach disk only when you export them.
- No credentials stored. The session comes from the Salesforce CLI (
sf org display,
sf org auth show-access-token), lives in memory for the run, and is never logged or written.
The CLI's telemetry is turned off for the calls the extension makes.
- Salesforce domains only. The session is only sent to
*.salesforce.com, *.force.com and the
other Salesforce domains; any other host is refused and redirects are not followed.
- No telemetry. The log (Migration Validator: Show log) contains counts and timings, never
record values.
- The extension remembers your last mapping per object in the workspace (column and field names,
transformations) so re-runs are quick. It never stores record values.
Settings
| Setting |
Default |
Description |
migrationValidator.cliPath |
(empty) |
Path to sf if it isn't on your PATH (machine-level setting only) |
migrationValidator.language |
auto |
auto, en or es |
migrationValidator.queryMode |
auto |
auto, rest or bulk (can also be changed per run in the editor) |
migrationValidator.bulkThreshold |
10000 |
In auto, records from which Bulk API 2.0 is used |
migrationValidator.sourceEncoding |
auto |
auto (UTF-8, falls back to Windows-1252), utf8, windows1252 |
migrationValidator.sourceTimezone |
utc |
Time zone of source date/times without offset: utc or local |
migrationValidator.maxIssuesStored |
50000 |
Discrepancies kept in the report (counts are always exact) |
migrationValidator.requestTimeoutSeconds |
120 |
Timeout per request |
Tips and limits (v1)
- The matching key matters most. Best: an external ID / legacy id field in Salesforce. Without
one, use a composite key (e.g.
Email + LastName); keys must be unique on both sides or the
report flags duplicates.
- Use the org filter if the object already had records before the migration; otherwise they show
up as extra.
- One object per run and one relationship level (
Account.Legacy_Id__c, not Account.Owner.Name).
- Compound fields (addresses, geolocation) are compared through their components
(
MailingStreet, MailingCity…).
- The whole source file is read into memory (up to 512 MB).
Requirements
- VS Code 1.90 or later.
- Salesforce CLI (
sf) with the target org authenticated, and a user with API Enabled and read
access to the object and fields.
Español
¿Se migró todo bien? Compara tu export de origen (CSV o JSON) con los registros de la org y
obtén qué falta, qué sobra, qué valores difieren, qué se truncó y qué lookups quedaron rotos.
Solo lectura y 100% local: los datos de la migración (a menudo PII de clientes) no salen de tu
equipo. Interfaz e informe en español (sigue el idioma de VS Code o el ajuste
migrationValidator.language).
Salesforce is a trademark of Salesforce, Inc. This extension is not affiliated with or endorsed by
Salesforce, Inc.
| |
