Analyzer tests

Analyzer test templates are the supported validation system for export attributes. You pick from a curated catalog of templates, supply a small JSON configuration, and the assignment runs against every row of the export.

Each template is parameterized and safe to expose to all authenticated users — no admin role required, and updates to the engine apply automatically without per-assignment migration.

You need a valid access token. See Authentication for how to obtain one with the Client Credentials flow. Pass it as a Bearer token in every request:

Each assignment runs independently against every row. A row fails validation on an attribute if any assignment on that attribute flags it.

If you need "value must be in range A AND format B", create one assignment for each constraint and the row fails if either fires. If you need disjunction ("matches pattern A OR pattern B"), express it inside a single regex pattern using regex alternation.

Validates numeric values against a threshold using comparison operators.

Use cases:

Configuration:

Example — validate price is at least 0:

Validates string length.

Use cases:

Configuration:

With skipEmpty: false, empty values have length() = 0, so length(empty) <= 100 is true (passes) and length(empty) >= 1 is false (fails). With skipEmpty: true, empty values always pass.

Example — title must be at most 100 characters:

Validates that a value is in (or not in) a predefined list.

Use cases:

Configuration:

Example — availability whitelist with case-insensitive matching:

Example — brand blacklist with placeholder values:

Validates a value against a regular expression.

Use cases:

Configuration:

Pattern delimiters

The pattern can be supplied with or without delimiters:

JSON escaping

Backslashes need to be escaped in JSON strings. When writing patterns:

skipEmpty and shouldMatch interaction

Setting skipEmpty: true is most useful when shouldMatch: true (i.e. "value must be a valid format"), since you want empty values to be handled by a not_empty_validation test rather than failing the format check too.

Example — date format validation:

Example — flag descriptions containing HTML tags:

shouldMatch: false means "valid when does NOT match". A description containing HTML triggers the test; clean text passes. The # delimiters avoid escaping the / that would otherwise be needed in a URL-style pattern.

Validates that a value is not empty or null.

Use cases:

Configuration:

No configuration parameters.

Example — title is required:

You can assign multiple templates to a single attribute. Each assignment is evaluated independently, and a row is flagged for that attribute if any assignment fires.

Validate price is between 0 and 999,999. Two assignments — a row fails if either bound is violated:

Date must be present AND match YYYY-MM-DD:

The format test sets skipEmpty: true so an empty value is flagged once (by the not_empty_validation test), not twice.

Title must be between 10 and 100 characters:

In addition to templateId and configuration, the following optional fields can be supplied when creating or updating an assignment:

See the API reference for complete error response schemas and examples.

All authenticated users can:

Begin with the minimum validation that catches obvious bad data, then add complexity.

Step 1 — validate price is non-negative:

Step 2 — add an upper bound:

If a value-format test (e.g. "must match URL pattern") should not also flag empty values — because a not_empty_validation test already covers them — set skipEmpty: true:

This avoids the same row failing twice (once for being empty, once for not matching the URL pattern), which would distort readiness math without changing what gets flagged.

If your source data has inconsistent capitalisation, enable case-insensitive matching once instead of listing every variant:

Each assignment runs independently and a row fails if any fires. Conflicting constraints fail every row.

Avoid:

Prefer:

If you need disjunction within a single value-format check (pattern A OR pattern B), express it inside one regex using regex alternation ((?:A|B)), not as separate assignments.

After assigning templates, run the analyzer against a representative sample of the export's data and check the readiness output. Aim for the same per-attribute readiness numbers you saw under V1 if you're migrating an existing V1 catalog.

For help, contact your Productsup Customer Success Manager or our Technical Support Team via support@productsup.com.