Append similar

JulieCulligan — Wed, 17 Feb 2021 14:13:28 GMT

Problem: Datasets have different domains that mean the same thing

Example: Street Suffix' can be provided in different formats when they need to be joined you have to substitute the data manually

Ie: Crt, Ct, Court

Solution: Have a thesaurus/catalog type option where if the data contains part of these suffix to be joined in the record it recognises its similar and joins

Re: Append similar

KoryKramer — Thu, 01 Apr 2021 17:24:58 GMT

A quick suggestion on the title of this one - it would better if it were "Join similar" rather than "Append similar"

I got some input from a colleague on this one, and something that could work in this situation would be to generate a translation table:

Oid, desc, code

1, court, 1

2, ct, 1

3, crt, 1

4, street, 2

5, st, 2

6, ave, 3

Etc

And now 3 tiers of joins.

Join Origin table (street suffix) to translation table ‘desc’ == join1

Join Destination table (street suffix) to translation table ‘desc’ == join2

Then join1.code == join2.code == Final result (hide the fields you don’t want to see).

Also if they are really different domains, then the key is the code stored in the data, not the description. If the codes are different, then you need the pattern described, if the codes are the same but the descriptions are different, then no problem the join will work at the db level.

Hope this helps!

idea Append similar in Data Management Ideas

Append similar

Re: Append similar