« first day (780 days earlier)      last day (4386 days later) » 
01:00 - 23:0023:00 - 00:00

Zane
Zane
23:11
Okay I have an awesome question for you guys.
How in the world do you get duplicate rows on an identity column?
swasheck
swasheck
@Zane how many? SET IDENTITY_INSERT ... ON
and someone forgot to reseed?
Zane
Zane
3
swasheck
swasheck
out of ...
Zane
Zane
Somehow a previous dba was able to strip replication remove the keys from that table. Insert dupes on an identity column and get replication back up.
100K plus.
Its the first three. @swasheck and here's the odd part. The whole row isn't duplicated just the identity column portion....
I thought it throws an error if you try to turn identity on with Dupes.
swasheck
swasheck
@Zane i dont think so ... not unless there's a unique constraint.
sqlfiddle.com/#!3/db875/1
Υπερκύβος
Υπερκύβος
23:29
@bluefeet MySQL has this syntax: SUBSTRING(str FROM pos FOR len)
And Postgres as well.
Zane
Zane
@swasheck there was but like I said he seems to have killed replication then pulled the unique off then did this.
swasheck
swasheck
shrug
Υπερκύβος
Υπερκύβος
@swasheck is correct @Zane. Identity column can have duplicates.
Zane
Zane
which means the key is useless to me and is making it hard to pull from the source into a table where that identity column is the key.
Υπερκύβος
Υπερκύβος
Remove the duplicates (if they are only 3 rows, it won't be hard) and then add a unique constraint.
swasheck
swasheck
23:44
@Zane what is the key? is it a unique key? a clustering key?
Zane
Zane
@swasheck at the client site yes.
swasheck
swasheck
@Zane just the identity?
Zane
Zane
yup
Now when I pull into my DataStore it's the companyID storeID and then that identity value. Essentially these dupes make it impossible to link that credit data to a transaction and therefore these records are useless.
bluefeet
bluefeet
@Υπερκύβος that explains it. thanks
Zane
Zane
Alright well thanks crew catch you on Monday.
01:00 - 23:0023:00 - 00:00

« first day (780 days earlier)      last day (4386 days later) »