I'm not sure how big the files will be. One patient is only about 2kb and the way it works is that the service checks for any new charges that have been generated in a parent application and creates the file to be processed by another application. It cycles about every 5 seconds, so I wouldn't expect many patients per file. I should have said that there could be multiple patients earlier, so there could be groups of 7 identifier lines. As I hit each patient, I would append that identifier to the appropriate cursor. That's my thought now anyway. I fogot about the delimiter flag in the APPEND command. Thanks!

>How big is file? Is order of records significant? Any other details?
>
>BTW, you can specify pipe as field delimiter in APPEND FROM command instead of converting it.
>
>>I have a data stream like below and I want to "walk" through it to parse out the different pieces. There are 7 identifiers that start each line, but in the example below FT1 and DG1 have wrapped; not normally the case. Without getting too deep into the various parts that are not actually field elements, I would like to work through the text and pull out the data to populate a table. It's a lot like a comma-delimieted file, but not quite. I did think about having different cursors for each of the 7 identifier lines, tranforming the "|" into a comma, and just doing an append - and maybe that is the best way - but wanted to see what others thought. Thanks!
>>
><snip>