Skip to content

Validation

Every record is checked against its survey type by the same engine, whether it was typed in, imported from a file, or collected in the field. That is what keeps your data consistent: a record that is valid one way is valid every way.

In order, for each record:

  1. Coercion. Values are read into the right shape: text trimmed, numbers parsed, dates read as real calendar dates. A value already in the right shape passes through unchanged.
  2. Visibility. The “show this field when” conditions are evaluated, so the record only has to satisfy the fields that actually apply to it. Hidden fields are set aside.
  3. Required. A visible required field must have a value. (A field can also be required only under a condition; see Beyond the builder.)
  4. Type rules. Each field’s own rules run: length and pattern for text, range and decimals for numbers, real calendar dates, a known option for a choice, a valid geometry, and so on.
  5. Repeat groups. The entry count is checked, then every entry is validated on its own.
  6. References. Species and lookup values are checked against your organization’s lists, and attachments against storage.

Problems come back as errors or warnings. An error means the record is not valid and will not save as-is (a missing required value, a number out of range, an unknown option). A warning flags something worth a look without blocking the record (for example, a date in the future). On import, the preview lists the specific problem on each row so you can fix the file or the mapping and try again.