January 14, 2016 at 8:45 am
I have a partition that has... exceeded it's time line. The last file was created in May of 2015. I have a lot of data sitting in that final file.
I understand I have a few options. I can just act like that final file is the old world and start anew, I can try and move all that data around and flatten it back out, or I can sit here and weep silently.
I'm new to partitions. I didn't create this one even, but I understand what it's doing. It's built on a range right and done by date. The extra file that's normally at the end is large. These are wide tables, 180 ish columns. The normal month holds around 5 million records. The current final file has 45 million records.
I will say that right now everything works and we're not having a problem, but I'd like to get this cleaned up sooner than later.
I have the new files created. My biggest question is which option is the best/safest (from your perspective) and some advice on how to tackle this beast. I'm guessing the fastest way would be to do a 2016, new file new year idea.
Thanks in advance!
.
January 14, 2016 at 4:55 pm
I'd say the first thing to do is look for a proc or a job that does a monthly handling of the partitioning of the table. The person that created the partitions might have built one an it might be looking for.
--Jeff Moden
Change is inevitable... Change for the better is not.
January 15, 2016 at 7:14 am
There is no job that handles the monthly partitioning currently. I'm planning on working on that after I sort the partitions out.
.
January 15, 2016 at 8:06 am
In that case, I'd first make a backup of the database, switch out the large partition, and hammer one partition of data back into place at a time using CREATE TABLE, INSERT, and SWITCH "in", etc. It'll take a little planning on your part but you'll learn a hell of a lot about partitioning in the process.
--Jeff Moden
Change is inevitable... Change for the better is not.
Viewing 4 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply