Hi all, I have a quick question which I will post . If I am creating a hold file on the server as in “APP HOLD ADVANCEMENT” and choose to hold as an unstyled format, lests say say FOCUS is that my most efficient format ? I know I can index with FOCUS but I sees the creation of an XML file in the OTHER folder at the same time. I am looking at efficiency for both size of the file created and more importantly for best response time reading that data. We would like to create some datamarts as we lack a data warehouse at this time and this would be a critical decision

This message has been edited. Last edited by: Kerry,

If you're looking for efficiency at dealing with data sitting 'outside' of a database, I would look into FOCUS or even better, XFOCUS.

Now, this really depends on how you'll be accessing the data in that file. If it's random access (by key), then you must have an index and definitely FOCUS/XFOCUS is the way to go.

But if you'll read in sequential access from top to bottom, I don't know if FOCUS/XFOCUS will be significantly more efficient that a good old fix-length ALPHA hold file.

You'll obviously have to try different approaches and time your results to see which one behaves better for what you actually need.

You also need to keep, in mind that if you are joining to one of these files, you will need an index, otherwise you will have to cater for a nonindexed file.


Also, if space and load time is not that much of an issue, then index all columns and get yourself a Cube.....