ansaurus

Question

DOS Database - help needed to recognize what is it

Answer 1

+1 A:

Most of those older flat-file apps used proprietary(ie, non-standard) formats. If the db is a standard format, you should see some kind of identifier close to the header that tells you what it is.

If you can't determine the format by visually inspecting the file in a hex editor, your best bet is to trace through the C code that reads each record and reverse-engineer the format.

David Lively 2009-11-13 22:07:17

I was looking at a dBase file dumper that I wrote in the 1980s. I expected to see "DBF" reserved in the header somewhere. There's no such thing. It just begins with version number and then goes into the last update timestamp, number of records, record length, etc.

wallyk 2009-11-13 22:11:42

Ah, dBase. And sweet memories of Paradox... (shiver)

David Lively 2009-11-13 22:38:52

Answer 2

+2 A:

Almost every version of Unix including linux and Mac OS has a command called "file" that recognizes a huge range of file types by their content. Try copying one of the data files to a Mac OS or Linux computer and running

file [filename]

from the command line.

rjmunro 2009-11-13 22:10:07

That's well worth the trouble to try, but now that I reflect on it, there were few patterns present in files of that era that the file command could know about, let alone inspect. Recognizing an old MSDOS .COM file in this way is nearly impossible. It just begins with instructions—no header, no containers—nothing. I believe the file extension had great importance for declaring the file type in those days.

wallyk 2009-11-13 22:14:44

I'll give it a try, but that might not help much, because those extensions are probably in polish "baz" which might mean "baza" == "database" and ind - "indeks" == "index"

kubal5003 2009-11-13 22:37:16

Tried it on Debian, "file" doesn't recognize it.

NXT 2009-11-13 22:55:05

I was going to suggest that if you try the Unix `file` command, you might also try `strings`, but if the file contents are Polish, the strings might resemble random sequences of consonants. (Well, that's how it looks to me ;-)

pavium 2009-11-13 23:28:28

Answer 3

+1 A:

Sounds like a dBase file to me. They were very common. It's not necessary for DBF to appear in the header. See the format description here:

http://www.dbase.com/knowledgebase/int/db7%5Ffile%5Ffmt.htm

edit Better link:

http://www.clicketyclick.dk/databases/xbase/format/

What's the value of the first byte?

I just double checked some DBF files that I have on hand and they do not have DBF in the header.

NXT 2009-11-13 22:21:46

Not dBase unfortunately, I would recognize it immediately and that was my first thought.

kubal5003 2009-11-13 22:28:54

I've investigated the link that you posted and that might really be it. I was thinking about the dBase that I know from windows, but I see that there was much more than that.

kubal5003 2009-11-13 22:45:12

I don't think it's dBase. I posted another link above that is better.Fortunately it looks very easy to write a decoder in C. Hints: Byte 2 (0x52) looks like the record length-2 and I'm guessing that 0x90 0xAD is the record separator.

NXT 2009-11-13 22:50:38

ansaurus

tags:

views:

answers:

DOS Database - help needed to recognize what is it

related questions