Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Hello All,
We have a large data set, approximately 2 billion records in a facts table that will continue to grow overtime. The Facts table will be updated weekly by adding last week's records and removing the oldest week in the process. Most entries in the community related to performance indicate using Qlik generated IDs is the best way to go, reduces storage needs and improves performance. However, a common problem listed when using AutoNumber() to generate these IDs is that the AutoNumber() function restarts the sequencing whenever a reload is performed. Does anyone happen to know if AutoNumberHash128() or AutoNumberHash256() ensure uniqueness across multiple reloads. We would prefer not to have to generate new IDs using AutoNumber() for the entire dataset each time we perform a reload.
I thank you in advance for you time and insight.
Best,
Carlos
don't think hash (for hash collision), look at Gysbert answer here
maybe an idea could be to store the last id and
rowno() + last id
You can create persistent keys with the Hash* functions. Leave off the Autonumber() part.