INPUT statement treatment of "accented" ASCII chars

[dmorrell]

Reading a flat text file in ASCII, fixed length records, fixed-width fields. One field is department name and the word "Café" appears in it. On the records where this happens, INPUT pointer jumps backwards 1 byte and therefore my @xx specs are all off for the remainder of the record and the data gets hosed.

Is there an INFILE option or parm that I can use to keep everything lined up?

thanks

Doug

[Patrick · Accepted solution]

Are you sure your text file is ASCII and not ANSI or UTF? A good way to check is opening the file with Notepad++ and there check in the encoding menu what it tells you.

Not sure why your pointer would be "jumping backwards" but I sure had already my fun with the "é". There is an "encoding" option available for the infile statement which you should set to the value of your source file.

[dmorrell]

To be more specific, the pointer isn't "jumping backwards". I'm trying to read the field at position 100 (@100) for example, but when there is "Café" in the preceding field, position 100 is blank and the data is actually at 101.

I always assume I'm dealing with ASCII when reading text files on a PC, and the file is sourced from USA. So I downloaded Notepad++, found the encoding was UTF-8, put that option on the INFILE, and it works fine now!

Thanks for the tip!

Doug

[Patrick]

The encoding of the little e with accent uses obviously 2 bytes but you're only reading the first byte with ASCII (and then you get one byte "behind").