At 08:20 AM 7/12/2005, Peck, Jon wrote:

>The compute below produces the location of the first digit, returning >0 is there is none. Note that the third argument of index, 1 in this >case, causes the search string to be considered as n strings of length >1 rather than looking for the pattern 0123456789. > >compute DigitLoc = index('0123456789', name, 1).

Thank you for reminding us! This is really what the third argument of INDEX is for. (Now, if a *negative* third argument meant "search for the first character NOT on the list", we'd really have something.)

But, should it be

compute DigitLoc = index(name, '0123456789', 1).

or am I getting confused from the sun? Syntax manual:

INDEX(a1, a2, a3) Return a number that indicates the position of the first occurrence of a2 in a1. a1 is the string that is searched. a2 is the string variable or string that is used in the search. If a3 is not specified, all of a2 is used. For example, INDEX(ALPHA8,'**X*') returns 2 for a case with value X**X**X* for variable ALPHA8. The optional a3 is the number of characters used to divide a2 into separate strings. Each substring is used for searching and the function returns the first occurrence of any of the substrings. With the same value X**X**X* for ALPHA8, both INDEX(ALPHA8, '**X*', 2) and INDEX(ALPHA8, '**X*', 1) return 1. a3 must be a positive integer and must divide evenly into the length of a2. The target variable must be numeric. If a2 is not found within a1, the value 0 is returned.