The object of the exercise was to produce a program for an EPOC32 machine
to convert Psion Series 3x OPL database files into their OPL32 equivalent.
It does not produce files for the Series 5/Geofox Data applications.
These programs are really meant for database programmers who wish to
convert a Psion Series 3x database application into a Series 5 or Geofox
version.

The program CONV_DBF.OPO was designed to read a straight .ODB file produced
by an EPOC16 OPL program. It should also read files from the EPOC16 DATA
application. Initially the application will list the total number of fields
in the database, and the types (string, integer, long, and float), and then
process each record. A text file is produced in the form:

"Apples","","3.6","10779","",""
"Prunes","Dried","4.2","23119","",""

The above example, there are six fields, the values are surrounded by
double quotes ("), and the fields delimited by commas.
In the case of OPL database files, all fields are shown, even if they
contain no data. For DATA files, which have 32 string fields by default,
all empty fields after the last non-empty field are omitted.

There are two ways this file can be used. Firstly the file is suitable
for importing into OPL32 DATA (File->More->Import text file), as it conforms
to DATA's default import layout. You need to create an empty DATA file with
field types and names to suit the file. In the above example, all fields might
be strings, or the third field could be a float, the fourth a number, etc.

Secondly the files could be converted into an OPL32 database file. I have
included a converter for this purpose, text2odb32.opo, which reads both
the binary log file and the CSV text file produced by conv_dbf.opo.

The main CONV_DBF.OPO program should be copied to a suitable folder on your
Psion 5 or Geofox. It will initially look in the same folder for OPL16
database files, so you can put them there as well. Series 3x database files
should be copied to this folder - there is no need to try any sort of file
conversion before or during the copy (e.g. with PsiWin).

The dialog produced on running the OPO will have a file selector, and at the
bottom, a choice field: View file or Write CSV file. View will allow you to 
examine each record without writing any files, Write will process each record 
without stopping.

After making your choices, you will first see a list showing the type of
each field in the database. For EPOC16 DATA files, there will normally be
32 string fields, although not all may be used. For OPL database files,
the actual number of fields will be shown - which can be either string,
integer, long integer, or floating number.

Pressing Esc will abort here, or any other key will read through the data
a record at a time, step by step (View), or continuous (Write), as selected 
earlier. In the View mode, you can stop by pressing Esc at any time.
If the records are large, they might overflow the screen.
At the end of the file, a short summary is shown, which lists the total
number of records found, and the highest number field used. This data is
written to a *.log file.

TEXT2ODB32.OPO

This file should be copied to the same folder as CONV_DBF.OPO.

I am grateful to Otfried Cheong for his OPLDB1.OPX which this program uses.

On running the program, a dialog will request the name of the *.log file.
Only these log files are shown in the file selector.

Next, you will be prompted for a table name. The default is "table1", but
you may enter any name up to 12 characters.

Next, you will be asked for a name for each field. A short name is advisable
as there is an overall limit of about 230 characters for the total list of 
names.

Note the table and field names you enter. These will be needed when you
write code to use the new database, although you may be able to read the
names with Otfried Cheong's OPX.

There are a couple of limitations to this program.
The size of each string field will be limited to the maximum found for each
field in the initial OPL16 database. This means that if you write any more
records to the new OPL32 database, you cannot write string data above these
limits. 

Furthermore, the new file will have only those fields actually used in the
original. So if you start with a 32 field Series 3a Data file, and only
the first 10 fields are used, then the new file will have only 10 fields.

The limits are stored in the *.log file, so I have included a small
program EDLOG.OPO which allows you to increase the size of the text fields,
and add new fields.

Note that this program may crash if the log and CSV files are invalid.

The output file name will have the extension .odb32 and so you may need to
rename your file to something sensible.

FILES

The files in the archive should be:
CONV_DBF.OPO    the program file.
TESTDB.ODB      an example OPL16 database file.
TEXT2ODB32.OPO  which reads the CSV and log files, and writes an OPL32
                database file.
EDLOG.OPO		a log file editor.

All the above files should be copied to the same folder. It is a good idea
to create a folder for you file conversions, where all files used and
produced will be kept.

OPLDB1.OPX      Copy this to c:\system\opx\


LIMITATIONS

These limitation apply to EPOC16 DATA files:

Only text fields of less than 256 characters are handled.

Addresses from a Series 3x DATA files will separate into one field per line
of orginal address. This may result in data in the wrong fields! These 
programs are really intended for use on OPL16 database files!

Only type 1 records are handled (these are normal records).

Field labels will be lost.

DISCLAIMER

These program are fairly rudimentary, and not guaranteed to be suitable
for any task. The user installs and uses the programs at her/his own risk.
The author will not be held responsible for loss of data, or any ill effects
on any computer it is held or used on.

I apologise for any mistakes in this readme file, hopefully the programs
will be self explanatory.

SOURCE CODE

The source code may be obtained by email from the author.

COPYRIGHT

OPLDB1.OPX is Copyright Otfried Cheong and RMR Software 1998. Without this file, 
it would not be possible to create TEXT2OPD32.OPO.

CONV_DBF.OPO, Edlog.opo, and TEXT2ODB32.OPO are Copyright Roger Muggleton 1998.
You may use and distribute the files in the archive freely.

COMMENTS

Comments etc. to the author:
Roger Muggleton
email: hzk@cix.co.uk