VERSIONS: P7
SUBJECT: Importing wide text files
BY: QUENTIN CORRELL
Date: 11 July 1997
I regularly import fixed ascii files of over 2500 bytes per record. In arriving at a stable app here are some of the things I found:
- If the next to last field in a record is a number field it may bomb. I had to import it as an alpha field and convert it later.
- If the records are not exactly all the correct length the import will fail. There is a bug in every UNIX fixed ascii output routine I've ever
encountered that if the last field is blank the record will NOT be padded to the correct fixed record length!!!
I use a textStream routine to read and copy all fixed ascii files and expand short records to the correct record size!
- Do *NOT* import fields out of ordinal sequence!!! *Especially* near the end of a record!