At 12:10 PM 7/08/2005 +0600, you wrote:

>Hello Helen,
>
>HB> The way UNION is currently implemented, the structure of the output of the
>HB> *first* SELECT statement sets up the structure for the all of the
>HB> succeeding SELECTs. Your problem here is that the first SELECT
>produces an
>HB> INTEREST_NAME column that is VARCHAR(6), while the succeeding one will
>HB> output INTEREST_NAME values that will overflow a VARCHAR(6).
>
>Let's suppose that INTEREST_NAME is defined as VARCHAR(30) in a DB.
>Could you clarify please why filtering the field content should
>"override"/substitute the field definition with VARCHAR(6) in the
>output structure? Does the usage of 'SELECT *...' vs 'SELECT
>INTEREST_NAME...' matter for the "overriding"?

Thinking about it later in the day....no, it shouldn't. The first SELECT
ought to define the output as <however it is defined>. It's only going to
be a problem with derived fields. Possibly "SELECT * " in the problem
description is disguising the actual query that is raising the exceptions.

./heLen