Could be. I ran the current production process (non-binary index of deleted()) against the same csv file last night without incident.

The next time that I run the process, I may just add some code to capture the cursor to a table and then add the binary deleted() tag from the command line. For some reason, I was under the impression that the binary index is smaller that a conventional structural index, but that may not be the case.


>>I have a process that loads about 8 million rows of data from a csv file, massages the data, and subsequently inserts the data into a SQL server table. In an effort to improve performance, I changed an index tag that I build on DELETED() to binary.
>>
>>The entire process ordinarily takes around four hours from start to finish. I launched it last night around 6:30 PM (against a test server) and found that it was still attempting to build the indices (DELETED() Binary, a Varchar (25), and a couple of integer keys at 9:30 this morning. The cdx in the tmpfiles directory appeared to be a 0 byte file. If memory serves, the cdx is usually several hundred mb.
>>
>>Obviously, I will retain the regular index on DELETED(). Just wondered if anyone else has run into this.
>
>Haven't experienced it, but it sounds like your index file may have crossed the 2GB file size limit. Maybe you can try creating the binary index into a separate .CDX file?