« first day (653 days earlier)      last day (1449 days later) » 

Martin
Martin
12:02 AM
0
A: SEDE refresh seems to be stuck (again)

DaniilSEDE now seems to be updated. Running this query results in: ... StackExchange.Drones.Meta 2020-05-10 12:29:43 StackExchange.Drones 2020-05-10 12:29:45 StackExchange.Or.Meta 2020-05-10 12:29:47 StackExchange.Or 2020-05-10 12:29:50 StackExchange.Tezos.Meta 2020-05-10 12:29:52 StackE...

 
 
7 hours later…
rene
rene
7:24 AM
It either unstuck itself or the backup was stuck which caused the on-call SRE to be notified and with starting the backup so did the SEDE refresh start.
 
 
1 hour later…
Martin
Martin
8:47 AM
Some time ago I played a bit also with a similar query for slovakia: Top Users from Slovakia.
I wanted to ask whether there is a simple way to handle diacritics. I mean, can I somehow catch both Zilina and Žilina (similarly Kosice and Košice) without typing one condition with and one without diacritics?
 
rene
rene
9:32 AM
 
Martin
Martin
10:01 AM
Thanks. I have edited my query a bit, following your advice.
 
rene
rene
@Martin you don't need the lower statement either.
 
 
2 hours later…
Martin
Martin
11:40 AM
Yes, I have seen in the link you've posted here that CI = case-sensitive, AI = accent-insensitive
Anyway, in some cases I do not need removal of diacritics - such as Nitra.
So there I simply left lower(Location) LIKE '%nitra%'.
In this way it looks "more uniform". And I was too lazy to edit away all instance of lower that were there from the older version. (When I did not know that I can use COLLATE Latin1_General_CI_AI.)
 
 
3 hours later…
rene
rene
3:05 PM
Okay. Do know that lower() has a performance penalty (I assume using a colllation performs better but don't have actual stats). Mostly memory will suffer. Not a problem on not too large tables
 
Martin
Martin
3:19 PM
So as far as performance goes, I might be better using collate even for the entries with no accented characters?
 
rene
rene
3:41 PM
Yeah, if you want to apply the same trick to the Posts table body field (so many, many more rows with loads of text) then you are more likely to run into the timeout due to the overhead of lower(). It is a rather "expensive" operation.
 

« first day (653 days earlier)      last day (1449 days later) »