Finding non-identical records in two different files

[John May]

I have a 5,000 record file with 200 users based in eight locations accessing via browser. About 175 regular users update many fields of each of their 25-30 records one a record at a time, while the other 25 "specialusers" loop through and update a few fields of 25-30 records at a time. There are about 20 regular users and 3 "specialusers" at each "location". The fields accessed by the regular users and the "specialusers" are mutually exclusive.

At a given point in time "A", I have to print all the records and distribute them to each user. The regular users get a copy of each of their 25-30 records, while the "specialusers" get a list of only their fields for all the records at their "location", grouped by the names of the regular users at their location. All users get a couple of days to review the output and make corrections using the same process they edited with originally. At given point in time "B", I have to reprint only those records that have been modified since "A". This whole scenario happens three times per year.

I thought I could do this simply by having DateModified and TimeModified fields in the file. This didn't quite work, because the "specialusers" update loop caused every record in the group they were browsing to show a new date and time, even if only one record in the group actually changed. Modifying the code to have "specialusers" browse and modify only one record at a time is not an option.

I need a way to compare records in file copy "B" to the records in file copy "A" and identify those that are not an exact match.

I don't want to resort to making the first print at point "A" a "draft" and reprinting ALL the records at point "B", but that is my fallback position.

Any ideas?

[John May]

This is a web-based FM/Lasso system. I'm not a very experienced FM user but I was under the impression that it was not a good idea to try to run any kind of scripts from the web. Also, I need a solution that is passive, i.e., automated and invisible to the user.

I found in another post what may be an answer, written by Tom Parker, to this question. He writes:

=================================================

Create a new calculated field in each database called, for example, MatchField. Define it as a concatenation of all the fields in the databases you would like to compare:

MatchField = Name&Address&City&ZipCode&etc...(Make sure the calculation result is text.

Define this field in both databases.

Now, define a relationship based on a match between the MatchField in each file.

In the first file, create a calculation field called MatchChecker. Define it as Count(RelationshipName::MatchField). Any exact matches will return 1. Exact duplicates will return a number greater than 1. Nonmatches will return blank. You can perform a find on this MatchChecker field to determine which records don't match up. Enter Find mode, type in 1, check omit, and you get all your non-matches.

=================================================

Does this sound like a resonable solution???

[ January 07, 2002: Message edited by: John May ]

[Kurt Knippel]

User will need to enter "editing" mode via a script, then make changes and click another "save" button to exit editing mode. This save script would also then mark the record as edited by a normal user.

You can then do the same concept, but in an automatic way for the special users. In this case the record would be marked as edited by a special user only if the record was actually edited.

You will then basically ignore the Modification Date field, since as you said, all records will end up looking like they were modified. Instead you will simply look for records that were maked as edited by a normal or special user.