Desperate help needed for a legacy db

[TomBascom]

on the v.6 the bin directory not exist ...
you have to put the $DLC in the PATH and not the $DLC/bin

Correct. But apparently we've found a 7.2 directory now

... and if I well remember, the DataDictionary is a "prodict" binary.

Nope. The dictionary has always just been ordinary 4GL code. Starting with v7 (as best as i can remember) the dictionary code was placed in a prolib rather than in individual files.

You can get to it at startup via "-p dict.p" or by executing "dict" at the editor.

 

[Rayford Booth]

OK so i managed to get the .d files created. All my tables have been exported. BUT now how do i read the .d files?

 

[rstanciu]

as a CVS files ... all strings are under "" and the fields are sepatered by a blank space.

 

[Rayford Booth]

ya got that. Thing i dont like is that is does not add the headers. needed to get that from another file and copy and paste

 

[TomBascom]

The dump does not provide any headers. If you can compile code you can add the headers like so:

for each _File no-lock where _File._Hidden = no:
  output to value( _Dump-name + ".hdr" ).
  for each _Field no-lock of _File:
    put unformatted _Field-Name + " ".
  end.
  put skip.
  output close.
  /* this next line adds the header to the top of the .d file and creates a .dat file... */
  unix value( "cat " + _Dump-Name + ".d " + _Dump-Name + ".hdr > " + _Dump-Name + ".dat" ).
end.

Depending on Progress version there may also be a trailer on each file with information about file size, code page and some other junk.

The exported data is space separated with quoted strings. For instance:

"abc" 123 "This is a test string" 99.44 12/7/41

This is 5 fields of data. Dates are sometimes problematic as, depending on Progress version and various startup parameters, the century portion of the year may be assumed. (Progress uses a 100 year window based on the -yy parameter to specify the start year of the window.)

 

[Rayford Booth]

Thanks for all the info. Appreaciate the info.

What is the best ODBC data connector for progress?

 

[TomBascom]

Version 7 is hopelessly ancient, obsolete and unsupported. There is no ODBC for v7.

Upgrade to version 10 and you can use Data Direct drivers.

Technically, the upgrade is drop-dead easy for a UNIX character based application. But unless the business has been paying maintenance for the last 15 years (and neglecting to upgrade along the way) it will be costly.

 

[Rayford Booth]

Actually the system is running on a OLD HP computer running progress 6.3E03 (just confirmed the version being used for the active program). Even if maint was being paid, the system could not handle the upgrade.

Non of the Data direct stuff is backwards compatible?

 

[TomBascom]

ODBC didn't exist back then.

 

[StuartT]

ODBC did exist back then though not in the form it takes now as I remember attending a seminar where a London Based consultancy firm demonstrated conecting to a progress DB on a unix server with a very early version of Excel. It was using 6.3D and was around 1994

 

[TomBascom]

Mea Culpa. Let me rephrase... "ODBC didn't exist in a useful form back then. And there is no hope whatsoever of getting it to work with anything resembling a modern ODBC driver".

 

[StuartT]

Mea Culpa. Let me rephrase... "ODBC didn't exist in a useful form back then. And there is no hope whatsoever of getting it to work with anything resembling a modern ODBC driver".

Agreed, no hope whatsoever using ODBC, all he can do is work with the raw data files and get them loaded into a more modern DB

 

[Rayford Booth]

I will look into updating the progress software. BUT i am going to have to clone the HD to a new machine.

Know of a good HD cloner that i can run from windows to clone the unix server over the network???

 

[oldemanw]

Hi Rayford,

I don't know about any Windows tools that will copy a HP disk remotely.
If you just want to keep the data, look at Tom's post about getting headers. He's giving very valuable hints.

However.
Using the "dd" command on the HP box, you can copy the contents of the harddisk to file, then copy that file (from the server) across using "scp" or a similar tool to transfer across the network.
On a Unix/Windows station, the file could be transferred to a harddisk (or partition) again, leaving you with a copy and data that is accessible.

If I recall correctly, your database is on a HP server on Unix (HPUX?).
You may want to get hold of someone who is comfortable with that OS to do the copying for you.
Or get acquainted yourself, but that takes a lot of man-page reading. I'm afraid you might get lost in the woods, learning Progress and Unix at the same time. I don't think it can be done in just 2 weeks. (No offense meant, it just takes time to sink in)

Willem

 

[Rayford Booth]

The thing is that i do not have any disks or anything. So i want a exact duplicate so if the HD crashes i can load on a new hd and start where i last backed up.

I will keep looking.

Thanks