Question for devs who work with databases. Do you constantly get tasks that involve joining data not on keys? Data from Exel files, other servers, etc where you have to try to match things on dates, amounts and other normal data.

Yeah, all the time unfortunately - we often get handed all sorts of random spreadsheets and CSV files by clients that we have to try to match up with data we store, and to make things worse the data we need to match against may not even be stored in a structured, atomic fashion (depending on what it is.)

I usually manipulate the file as a spreadsheet to just get the "columns" I want, export to CSV, then use a bunch of regexes to create the correct SQL statements I need. It's not pretty, nor is it quick, but it gets the job done.

From the DBA perspective, we tell them no, but we'll be more than glad to create a DB in SQL Server, Oracle, or Postgres, whichever fits what they are doing. I've ported many a spreadsheet or MS Access DB to one of those relational DBMS.

Bah.. I deal with these kind of shit on a daily basis.. And the worst part of it is that the data is from the actual database. Fucking composite keys.. I hate them with a passion

@imerljak of course.. Composite keys can be good (it is said..). The problem is bad design. Really really bad design

Recently started a job and I get these types of tasks multiple times a week. The worst is when the data is from a different system the customer has and I'm matching on first and last names, but the names are slightly different in both systems (Andy for Andrew) and the customer asks why I couldn't tell these were the same people when comparing this through queries.