Ever had a database that suddenly and unexpected grow very much? If it happens during…
Alright, so you get this error when you process your dimension. You look carefully throug your data looking for duplicate values, and you find NOTHING!?! Recognize the situation?
I had a frustrating time last week trying to understand the reason for this (fairly straight forward) error message. I looked very carefully at the data rows and could see duplicates in my SELECT…GROUP BY statement, how is that even possible?! Well it turns out that the data set was in really poor condition. The strings (addresses from a web form) were not cleaned from strange characters, so I had strings with leading and following characters in tha range from 0x01 to 0x31 (mostly non printables; look in an ascii table), thank God for CAST(SUBSTRING(address,1,1) AS BINARY(2)) AND CAST(SUBSTRING(REVERSE(address),1,1 AS BINARY(2)). Now I am going to have a little chat with the web developers on how to verify web form data BEFORE shipping it off into the backend systems…
//Joakim