End-of-file Problems

[Eric Steinmann]

Hey everyone,
I'm having an interesting issue insheeting a very large .csv dataset. One of the fields consists of a series of random characters which sporadically contain DOS end-of-file markers. These EOF markers inevitably stop the import leading to truncated data. If I could open the .csv file I could replace the EOF characters with something else, but it's too big to open in any program that I've tried. I know in SAS there is an option to ignore the EOF markers (IGNOREDOSEOF). Is there anything like this in Stata? If I had SAS I could import, then just export a .dta, but I don't. Any ideas?

Any help is appreciated. Thanks!

[Sergiy Radyakin]

Use filefilter to replace those characters with something else, import and replace back. For example if the # sign is never used, you can temporarily use it as a placeholder for EOFs.
Sergiy

[Rich Goldstein]

to expand slightly on Sergiy's response: I would start with hexdump (use either the analyze (my preference) or the tabulate options) to see what is never used; if it is something "odd", you can download (findit using search) the asciiplot command to get the right code

[Eric Steinmann]

I appreciate the input guys! Yes, I found a solution following your suggestions. Saved me!