Re: bug in substring???

scott.marlowe wrote:
> gives me nothing.  Shouldn't a negative offset, or even 0 offset result in 
> an error or something here?  Or is there a special meaning to a negative 
> offset I'm not getting?

In varlena.c there is this comment:
 * text_substr() * Return a substring starting at the specified position. * - thomas 1997-12-31 * * Input: *    -
string*    - starting position (is one-based) *    - string length * * If the starting position is zero or less, then
returnfrom the start *  of the string adjusting the length to be consistent with the *  "negative start" per SQL92. If
thelength is less than zero, return *  the remaining string.
 

Joe