data
Use data for portable export, validation, and import workflows that sit above the raw Table API.
snow-cli data <verb> [options]
All data subcommands also accept the global flags from the command overview.
data export <table>
Export records from a single table into a portable artifact.
snow-cli data export <table> [options]
Important options:
--query <encoded-query>: restrict exported rows--fields <a,b,c>: only export selected fields--limit <n>: maximum records to export--order-by <field>: sort the export-o, --out <path>: write to a file instead of stdout
Examples:
snow-cli data export incident --query 'active=true'
snow-cli data export incident --fields sys_id,number,short_description --limit 50
snow-cli --output csv data export sys_user --fields sys_id,user_name,email --out users.csv
Use this command when you want a flat, single-table artifact that can later be validated or re-imported.
Note: some exportable fields aren’t re-importable. sys_id can always be
exported, but data validate/data import reject it as not writable for
create-style imports (field_not_writable). If an artifact is meant to be
re-imported, export the writable fields you actually want to recreate rather
than a full dump that includes system-managed fields like sys_id.
data export-package --file <spec> --out-dir <dir>
Export a multi-table dataset package from a manifest spec.
snow-cli data export-package --file dataset-spec.json --out-dir exported-dataset
Important options:
-f, --file <path>: dataset export spec file--out-dir <dir>: destination directory for the manifest and exported table files
Examples:
snow-cli data export-package --file dataset-spec.json --out-dir exported-dataset
snow-cli data validate --file exported-dataset/manifest.json
snow-cli data import --file exported-dataset/manifest.json
Note: data export-package does not support --output csv.
data validate --file <file>
Validate an export artifact or dataset package against the target instance.
snow-cli data validate --file export.json
Important options:
-f, --file <path>: file to validate
The validation report checks schema compatibility and reports errors and warnings before you try an import.
Note: data validate does not support --output csv.
data import --file <file>
Import a flat export artifact or dataset package.
snow-cli data import --file export.json [options]
Important options:
-f, --file <path>: file to import--dry-run: preview the import plan without creating records--import-set-table <table>: use the Import Set API for a flat table-export artifact--fail-on-error: exit non-zero when Import Set responses contain row-level errors
Examples:
snow-cli data import --file users.json
snow-cli data import --file users.json --dry-run
snow-cli data import --file users.json --import-set-table imp_user
snow-cli data import --file users.json --import-set-table imp_user --fail-on-error
Notes:
--import-set-tablecurrently works only for flat table-export artifacts, not multi-table package imports.- When
--import-set-tableis not used, the CLI falls back to direct create-only Table API writes. --fail-on-erroris mainly useful together with--import-set-table.data importdoes not support--output csv.
Real-world export → validate → import workflow
# 1. Export records to a file
snow-cli data export incident \
--fields number,short_description,priority,state \
--limit 10 \
--out export-incidents.json
# 2. Validate the exported file against the target instance
snow-cli data validate --file export-incidents.json
# 3. Preview what an import would do (dry-run)
snow-cli data import --file export-incidents.json --dry-run
# 4. Import (creates new records via Table API)
snow-cli data import --file export-incidents.json
The validation step checks that the fields in the export exist on the target table and reports any mismatches before you attempt the import.
When to use data instead of table
Use data when you want:
- a portable export artifact,
- schema validation before import,
- a repeatable dataset package,
- a dry-run preview of an import,
- an Import Set-backed import path.
Use table for direct record-by-record CRUD operations.