如何在子字符串定义的SQL中识别和删除重复项

Question

I have a complex issue regarding de-duplication in SQL that I could use some advice on:

I have a table with airport codes. However, there are duplicates in some cases where one row lists the local airport ID, while another lists the ICAO (international) ID , which includes a leading K .

I need to identify duplicates such as the following: KI80 and I80 KX49 and X49

Note that there are many valid rows that start with a K .

Step 1: I need to identify the duplicates for the above cases.

Step 2: I need to use SQL to automatically delete all duplicates which have the leading K .

Step 3: I need to identify in a different table table b , which rows were using identifiers that I just deleted, so I can update them to the surviving ID (example: if they used KI80 , I need to change them to I80 in this new table")

Any help would be greatly appreciated!

Answer 1

You can use a self join in a delete statement. The idea is to join the table to itself, but doing the match on a "K" prefix. If a match exists, then the "K" record is a duplicate:

delete t
    from table t join
         table tnotk
         on t.airportID = concat('K', tnotk.airportID) and tnotk.airportID not like 'K%'
    where t.airportID like 'K%';

Note: this assumes that no non-ICAO airport ids start with a "K".