Observation Pointers / Table Vector after a REMOVE statement

[Ksharp]

But the condition is that your large dataset must be compressed firstly.

[data_null__]

The compress options renders the POINT option useless so how can this help?

[pre]
ERROR: The POINT= data set option is not valid for the data set WORK.ACCOUNTS, the data set must be accessible by observation number for POINT=processing.
[/pre]

[Ksharp]

Yes.
You are right.I overlooked this part in documentation.

Ksharp

[Peter_C]

PHK
the answer to the final question in your original post in this thread is probably NO.
Since you do not want to re-write the whole file when you "delete" a row these rows remain in the file and are marked as deleted. You seek to have a number running sequentially through the data skipping over and not counting deleted rows? Then you will need to maintain that column because SAS does not.
You seek a pointer into the data providing direct access by something like row number? Then this service needs to accept the presence of those rows you deleted ( but did not re-write the file to remove). This is the row number you want to consider "wrong"! (and what POINT= accesses)
I think you have to work with SAS as it is written - not the way you think it should!
You might find there is a flaw in your concept that requires the SAS approach - which I think
applies here.

Regards
peterC slp chkc
(and what POINT= accesses) edit by: Peter.C 14:00 BST