Informix Error -661

Reasons / Common Causes

-661 fires when a LET statement's left side (the variable list) has a different count than the right side (the value list) — per the official guidance, the two counts must match exactly.

1. A LET statement listing more variables than values, per the official guidance's example LET a,b = 10,20,39; reversed (too few values for the variable count is the mirror case).
2. A LET statement with multiple variables set from a single combined expression — per the official guidance's example LET i,j = proc1()+proc2();, where the right side evaluates to one value, not two, despite two functions being referenced.
3. A LET statement with multiple variables set from a subquery returning a single column — per the official guidance's example LET a,b = (SELECT c1 FROM tab), where the subquery returns one column's value, not two.

Solutions / Resolution

1. Match the number of expressions on both sides of the LET statement, per the official guidance.
2. For two functions combined with an operator, call them separately if two distinct values are needed, rather than combining them into one expression:

   LET i = proc1();
   LET j = proc2();
   

3. For a subquery, select as many columns as there are variables, if multiple values are genuinely needed:

   LET a, b = (SELECT c1, c2 FROM tab WHERE ...);
   

Examples

Mismatched literal counts

LET a, b = 10, 20, 39;
-- -661: 2 variables, 3 values

A combined single-value expression assigned to two variables

LET i, j = proc1() + proc2();
-- -661: the expression evaluates to one value, not two

Corrected — separate calls

LET i = proc1();
LET j = proc2();

Diagnostic Checks

1. Count the variables on the left of LET and the values (or columns) the right side actually produces, confirming they match exactly.

Related Errors / Related Topics

No closely related error codes are cross-referenced for -661 in this set yet.

Count both sides of the LET statement carefully — a combined expression or a single-column subquery produces one value, not one per variable named.