The main reason that ROW_NUMBER() and RANK() functions have not been added is that they require implementing the whole concept of windowing functions that are NOT part of the SQL standard that PROC SQL implements. SAS added the MONOTONIC() function BEFORE that definition of windowing functions for the SQL language was even created.   To implement ROW_NUMBER() SAS should create a flavor of SQL using a more recent standard.  They could either do it as a new PROC or perhaps as an option in PROC SQL.   ... View more

If you are using Teradata and need to convert an attribute that is a character-type to a numeric-type, let me suggest that you do the work using "explicit SQL" in the native Teradata SQL dilect, and not in SAS-SQL.   The example below shows the steps I use.  Only the last is truly needed.  The regex replace function removed any remaining non-numeric characters, like any potential nonprinting characters.   SELECT     '00000000012345678912'     , TRIM( LEADING '0' FROM '00000000012345678912' )     , REGEXP_REPLACE( TRIM( LEADING '0' FROM '00000000012345678912' ), '[^0-9]+','',1,0,'i')     , CAST( REGEXP_REPLACE( TRIM( LEADING '0' FROM '00000000012345678912' ), '[^0-9]+','',1,0,'i') AS BIGINT ) ;   You may be able to use the INTEGER datatype. Instead of a BIGINT.   The number in your example (12,345,678,912) is a BIGINT, a 64-bit (8-byte) signed integer. Range: -9,223,372,036,854,775,808 to 9,223,372,036,854,775,807 ... View more

If I understand your question correctly, you want to use SQL to create two data sets at the same time. SQL (including SAS' PROC SQL) can only create one table at a time. ... View more