Yes the data integrity is an issue - this is infact the purpose of this report to try to identify amonst teh 100's of thousands of records "what's missing"
No its not quite that there are three tables making a complete record, its just THREE systems all with similar/duplicated instances of the same (similar) data and identifying "what's the same" is a challenge for our users, so the solution is to get the system to do as much matching as it can, and present the "completed" record (along with the 3 source rows) to the user, so they can decide if the computer is actually "right".
Thus the "fuzzy matching" suggestion...
eg:
(row) ID Name
1 C1 NULL SMITH source1
2 NULL J SMITH source2
3 C1 John Smith source3
4 NULL M Smith source2
So, in this example I'd like the #2 record to be "aggregated" with the others, however because its a
loose match I WOULDN'T want #4 to be included, so I can't just grab Lname...
No comments:
Post a Comment