I'm trying to process header/detail lines in a flat file, something I've seen
a few examples of, but still have some questions. I've seen Allan Mitchell's
comments and an example on sqlis.com (thanks for the fine examples!), and
tried an example from a BOL post. Still, I have a question about approach.
If I have several header/detail formats AND want to allow users to specify
their own format (I'm building a front-end that creates packages from user
spec's), is there a problem with storing the row formats in metadata and
consuming that at package runtime, so that I only have 1 script component
with 1 input (flatfile) and perhaps x (max supported) outputs, and I
determine the format of the output (number and type) columns by reading my
metadata at package run time?
OR, should I create fixed output formats based on the metadata (output
column formats) when I generate the packages? Thus having 1-off script
components, rather than my own, metadata consuming component, preset
component?
>> Stay informed about: Flat File - Different row types - still confused