Oninit
Talk to an expert →

Informix Error -391

-391 Cannot insert a null into column column-name.

This statement tries to put a null value in the noted column. However, that column has been defined as NOT NULL. Roll back the current transaction. If this is a program, review the definition of the table, and change the program logic to not use null values for columns that cannot accept them.

Oninit® Troubleshooting Guidance

Reasons / Common Causes

-391 is the most direct, everyday NOT NULL violation: an INSERT (or UPDATE) explicitly supplies NULL for a column that's declared NOT NULL. Unlike -292 (an implied/omitted column defaulting to NULL), this is an explicit NULL value provided in the statement itself.

  1. An INSERT statement explicitly passing NULL for a NOT NULL column — the direct, most common cause.
  2. A parameterized statement/prepared statement binding a NULL host variable to a NOT NULL column, often because the application-side value was never actually populated.
  3. An UPDATE setting a NOT NULL column to NULL directly.
  4. Application logic that doesn't validate required fields before submitting the statement, letting an unset/empty value pass through as NULL.
Solutions / Resolution
  1. Roll back the current transaction, per the official guidance, since the insert/update didn't take effect.
  2. Modify the application logic to supply a valid non-null value for the column, or ensure required fields are validated before the statement is submitted.
  3. For parameterized statements, check the host variable's value immediately before binding it, to catch an unexpectedly NULL value earlier in the code path.
Examples
An explicit NULL rejected
INSERT INTO customers (customer_id, name, email)
    VALUES (501, 'Acme Corp', NULL);
-- -391: email is NOT NULL

Fix — supply a valid value:

INSERT INTO customers (customer_id, name, email)
    VALUES (501, 'Acme Corp', 'contact@acme.example');
Validating before submitting
-- In application code, validate that required fields are
-- populated before constructing the INSERT/UPDATE statement,
-- rather than relying on the database to catch it after the fact.
Diagnostic Checks
  1. Check which column the error names and confirm it's declared NOT NULL.
  2. Trace back to where the value became NULL in the application logic — an unset variable, a missing form field, a failed upstream lookup, and similar.
Related Errors / Related Topics
  • -292 — "An implied insert column column-name does not accept NULLs." The closely related sibling — an omitted column defaulting to NULL, rather than an explicitly supplied NULL value.
  • -386 — "Column contains null values." A related NOT NULL-enforcement error, at schema alteration time rather than insert/update time.

Trace the NULL value back to its source in the application — this is almost always a missing or unvalidated required field rather than a database-configuration issue.