Hi Michel,
You may be able to do this using a postprocessor script. In the 'Load Information' tab of the application, put in the name of the script which will modify the .DB temporary file that gets created after indexing has completed. I find awk works great for postprocessor scripts since the .DB file is a delimited columnar file. An example script that may work for your needs would be:
script.awk:
BEGIN {
FS="\t"
OFS="\t"
}
NR > 1 {
# assuming $2 is 'amendment date' and $1 is 'order date'
if ($2 == "")
$2 = $1
}
To see what the .DB temporary file looks like you can put in a fake postprocessor script name, like "abcd", and then run arsload against the file. The load will fail during postprocessing and the .DB file should be out there. You can test your awk script against the .DB file using:
awk -f script.awk nameofdbfile.DB