« first day (2668 days earlier)      last day (2190 days later) » 

 
5 hours later…
6:09 AM
@jadarnel27 Thanks for those links
 
6:53 AM
morning
 
7:04 AM
morning and evening
 
Morning
 
7:39 AM
morning
 
8:36 AM
which reminds me of similar disaster, a few years ago
Jul 7 '16 at 18:09, by ypercubeᵀᴹ
Jun 28 '12 at 18:53, by ConcernedOfTunbridgeWells
@swasheck RBS is a large bank here in the UK that outsourced a lot of its operations and has just had a very embarassing outage.
 
@ypercubeᵀᴹ We had a similar incident a few weeks back
the bank asked their customer to send their transfers in by mail, I am not kidding you
 
@TomV I believe you.
 
> In the latter case, customers were asked to mail their name, address, date of birth, account number and transfer details.
 
9:14 AM
That's our job http://www.commitstrip.com/2018/04/24/thats-our-job/ -- Sponsored by @OpenClassrooms who is hiring mentors for their great mentorship program ! => http://mentor-en.jobs.openclassrooms.com/
 
 
2 hours later…
10:49 AM
@MikaelEriksson Have you had any problems with inconsistent query results when running XQuery against an XML column (one that isn't changing)?
 
11:10 AM
@JamesL I have seen strange things happen with memory when using OPENXML but nothing out of the ordinary with the XML datatype.
 
11:35 AM
@MikaelEriksson Do you know much about how the hid (hidden?) attribute is used in seeks on primary XML indexes? The value of hid used in the index seeks is the only difference I've found so far between the version of a plan that retrieves data and the version that retrieves none.
hierarchy id, maybe
This stuff, I mean:
<StartRange ScanType="GE">
                                                            <RangeColumns>
                                                              <ColumnReference Database="[LPS]" Schema="[sys]" Table="[xml_index_nodes_2066106401_256000]" Alias="[key:2]" Column="hid"></ColumnReference>
                                                            </RangeColumns>
                                                            <RangeExpressions>
                                                              <ScalarOperator ScalarString="'ã€Á€'">
Sorry, maybe I should post this as an actual question now that I'm starting to get an idea of where the issue might be.
 
11:54 AM
@JamesL If I remember correctly hid is a representation of the path to the xml node stored reversed (Item#ParentItem#RootItem) so it can be used to seek on path expressions. Because it is stored reversed it can also be used to seek on descendant path expressions (//Item) using a "like" range seek.
 
Thanks
 
 
1 hour later…
1:21 PM
Well that was weird. It looks like it's intermittently searching for a (wrong) absolute path instead of a relative one.
I'm hoping the developers are interested in not using XML indexes anymore.
 
@Taryn looks like y'all hit an index fragmentation problem
 
@sp_BlitzErik not my fault
 
are the indexes rebuilt now?
@MaxVernon you didn't have any good competition here:
2
A: CPU clock speed versus CPU core count - higher GHz, or more cores for SQL Server?

sp_BlitzErikHold on hold on hold on While the performance and licensing aspects are interesting, they're not the only aspect of a workload to consider. One thing that can have an impact on processor choice is worker threads. Worker Threads? Yeah buddy! They're the things that your SQL Server will use to...

 
1:39 PM
@sp_BlitzErik nice one. Also, connection pooling.
 
sounds like developer talk
 
or should I say, badly written handling of open connections in lousily written client apps that use connection pooling
ask me how I know
 
i can probably guess
ahem
but i won't
 
@MikaelEriksson since you're here, and such a fantastic person, how do I write an xquery to see if an attribute value contains an "@" ?
@sp_BlitzErik my ego thanks you
30,000 user connections is a lot, is all I can say.
d'oh, Mikael is gone.
 
@MaxVernon whatever[contains(@attribute, "@")]
 
1:45 PM
I tried [@attribute][contains(., "@") = True], but that makes me look dumb.
@sp_BlitzErik of course. I'll give that a whirl
cheers
 
@MaxVernon not really gone, only busy doing boring work. Looks like Erik sorted things out for you.
 
@MikaelEriksson hopefully that works!
 
@MikaelEriksson yes, that seems to have almost worked. Like so many things in XML, it's a mater of tiny slices of my life gone whenever I try to figure out xquery syntax.
I guess my problem is I really need notcontains
 
@sp_BlitzErik shh
 
I'm trying to get sequences where a specific attribute doesn't have an "@" in the value.
 
1:52 PM
Well that’s it for me. Back tomorrow :)
 
:-D
 
@MaxVernon can you post sample xml in pastebin or something?
 
Apparently @Nick_Craver just shut down @StackOverflow rather than deal with the GDPR.
 
nailed it
 
2:07 PM
@sp_BlitzErik this should do the trick:
DECLARE @x XML = '<xml>
<customer name="Max"><email address="mvernon@mvct.com" /></customer>
<customer name="Erik"><email address="erik@your-mom.com" /></customer>
<customer name="Brent"><email address="brentcom" /></customer>
</xml>';
SELECT FullXML = @x
    , WithValidEmail = @x.query('/xml/customer/email[contains(@address, "@")]')
    , WithInvalidEmail = @x.query('/xml/customer/email[contains(@address, "@")] = False');
WithInvalidEmail returns "false" in the column instead of the actual sequence of <customer name="Brent"><email address="brentcom" /></customer>
 
@MaxVernon does that work how you want it or do you need help with something?
 
@sp_BlitzErik I'd like it to not return false, I'd like it to return the sequences where addresss attribute does not contain an "@" character.
ahhh, I think I figured it out.
WHERE @x.query('/xml/customer/email.exist('/xml/customer/email[contains(@address, "@")]') = 0;
thanks for your help, Erik - much appreicated
 
i think you can shorten it, hang on
 
0
Q: Editing other people's answers

Bertrand LeroyI'm new on this forum and I'm learning new things everyday (which is great!), and I am also learning about how to behave on this forum so here's a question about etiquette and all that, specifically about the etiquette regarding the editing of other people's responses. For example ... I posted a...

 
actually that doesn't really do it. I feel like I'm close, but stoopid.
 
2:16 PM
It should be moved to meta, where it can be closed as duplicated
 
@sp_BlitzErik I'm going to ask it on main, where people can really make fun of my skillz
 
0
Q: Editing other people's answers

Bertrand LeroyI'm new on this forum and I'm learning new things everyday (which is great!), and I am also learning about how to behave on this forum so here's a question about etiquette and all that, specifically about the etiquette regarding the editing of other people's responses. For example ... I posted a...

 
@MaxVernon lemme know when it's up, i have an answer
 
sweet, almost there
0
Q: Return xml sequences where an attribute doesn't contain a specific character

Max VernonConsider the following simple XML: <xml> <customer name="Max"> <email address="me@you.com" /> </customer> <customer name="Erik"> <email address="erik@your-mom.com" /> </customer> <customer name="Brent"> <email address="brentcom" /> </customer> </xml> I want to get a lis...

 
2:38 PM
@MaxVernon there you go
@MaxVernon you want to return it as xml or as the values inside?
 
@sp_BlitzErik I'd like the xml. Although, having the values of course is helpful too!
this makes me feel like a total nooob.
 
@McNets Moved and answered.
 
2:58 PM
@MaxVernon that should work
 
@sp_BlitzErik OK. .. - who knew!
 
yeah xml sucks
3
 
thanks very much
 
@PaulWhite Thank you
 
3:08 PM
@sp_BlitzErik Starred for the enduring truth of this statement.
 
Have any of you ran into a situation where OS patches fail and even Microsoft can't figure out why?
Because I can tell you it's very fun.
@sp_BlitzErik that is an easy way to earn some stars in this room.
 
@Zane taking the easy way out is my spclty
 
3:21 PM
@sp_BlitzErik is "woogy" the technical term?
 
@MaxVernon yeah it's short for "woah oh oh god why"
2
 
I'm sure when you see what I'm doing, you're going to laugh
@sp_BlitzErik that's some good material right there
lol:
Thanks for this info! However, FYI, I got this error when I tried it: "Msg 6335, Level 16, State 101, Line 2 XML datatype instance has too many levels of nested nodes. Maximum allowed depth is 128 levels. " Any thoughts on how to deal with that? — jrdevdba 1 min ago
a hand grenade comes to mind.
FYI, that is unrelated to my XML questions earlier.
 
@MaxVernon running on substandard caffeine is the equivalent to waking up drunk
 
@sp_BlitzErik without the fun the prior night, I take it. :-)
 
exactly
@MaxVernon query plans with > 128 nodes are stored in sys.dm_exec_text_query_plan rather than sys.dm_exec_query_plan
 
3:27 PM
SQL Server 2000 yikes....
I've said it before and I'll say it again. If your DBMS is old enough to buy you a pack of smokes it may be time to upgrade.
5
 
@Zane 100%
 
there was a sql skills email recently about a client problem they ran into with a customer who did an in place upgrade from 2005 to 2012 or something
IN PLACE
on 7 year old hardware
 
lord why? new hardware is going to be such a small portion of the cost of that upgrade.
and will make a massive difference!
even if you just ran it on a new laptop, it would be faster
 
@MaxVernon Because the moto of most companies is "Short term profits over long term stability"
 
@Zane yah, and they get what they get™
 
3:32 PM
I should know I work for Wells Fargo lol
 
lol
I wonder how many accounts they opened in your name, "for sales".
 
@MaxVernon my assumption is it's probably best not to try that on the IT guys.
Probably easier targets than me for this sort off thing.
 
lol. probably.
 
@BertrandLeroy Ha ha thank you. Not going to delete it (meta is for discussion so comments are rarely deleted here). — Paul White ♦ 58 mins ago
I am a hero, as I could resist no flagging this for mod with 'Except cases'
 
@MaxVernon dbfiddle.uk/…
 
3:51 PM
@McNets nice. It wouldn't hurt to add an answer with those variants!
 
Mr MD makes some horrible edits
 
@PaulWhite - have you seen this? It sounds like it might be something you'd know about.
 
@MaxVernon I'll do it at home tonight.
 
cool!
 
4:19 PM
1
A: Why would MS SQL server automatically remove manually created indexes?

Ken KimYou can setup a DDL trigger to capture create/drop index. CREATE TRIGGER [DDLAuditTrigger] ON ALL SERVER FOR CREATE_INDEX, DROP_INDEX Here is an article on howto. https://www.mssqltips.com/sqlservertip/2085/sql-server-ddl-triggers-to-track-all-database-changes/

Regarding this answer, where can we find what are the events that can be declared in a DDL trigger?
 
4:38 PM
@MaxVernon yes will answer later
 
cool
That's quite an explicit index name from Microsoft: IX_syspolicy_system_health_state_internal_target_query_expression_with_id
 
Professionals use a naming convention and spaces in the index name
 
good thing we're not professionals in here
 
Well I didn't say which profession
 
4:59 PM
what is a UDX plan operator?
> Extended Operators (UDX) implement one of many XQuery and XPath operations in SQL Server. All UDX operators are both logical and physical operators.
 
@MaxVernon CREATE TABLE dbo.[Your#Mom]
 
I knew it. If I see that in code someone dies.
 
@MaxVernon blitzcache checks for table scans as well
@MaxVernon ahem
 
> There's a number of poor choices that the vendor has made, this is a relatively minor one as far as things go.
:-D
@sp_BlitzErik I added those details to my answer, hope you don't mind!@
 
not at all
god, of all the things for a monitoring tool to warn you about, and you to focus in on
 
5:26 PM
@sp_BlitzErik I'm assuming you're talking about the OP?
 
I tested and confirm that by turning SERVER AUDIT off on 2016 SP1 CU8 and SP2, SET ALLOW_SNAPSHOT_ISOLATION ON worked as expected otherwise it would just hang. Thanks Tom V. — donger 20 mins ago
Glad that wasn't a comment answer @PaulWhite ^
 
@TomV that's a weird bug
 
@MaxVernon especially for a cu
 
@MaxVernon yeah
 
@sp_BlitzErik I provided that answer as part of my quest to provide answers for all the questions without answers.
I'm thinking it might take a while.
 
5:30 PM
@_@
 
and probably no one will be using it by the time I'm finished. Oh, who am I kidding, 2008 R2 will be in use until at least the 2030's.
 
that doesn't mean you have to support it
 
lol
we have some 2000 instances still hanging around here
granted they are on the way out, but you know.
 
@MaxVernon that's a lot of instances ;)
 
@MaxVernon When I read that at first, I thought you meant you had 2,000 instances of 2008 R2 :o
 
5:32 PM
@ypercubeᵀᴹ and @jadarnel27 nice one.
let me rephrase that; "we have around 8 SQL Server 2000 instances still hanging around."
ahhh the vagaries of the English language.
 
lol
Sorry, could not resist.
 
Wow it's quiet in here today.
 
hi James!
 
6:12 PM
People complaining about unresolved issues but refusing to log those issues in a ticketing system...
'we told somebody already and it still doesn't work'
Put the damn support email in cc and it'll create a ticket automatically, how hard can it be
 
The CC field is so far away though.
 
@jadarnel27 it seems it isn't if they want to complain and let managers know
 
@TomV LOL wow. That whole situation sounds quite frustrating.
 
@jadarnel27 oh does it show?
 
I'm just good at picking up on the subtle undertones of people's chat messages.
=)
 
6:32 PM
@TomV @TomV what ticketing system are you using? I've recently contracted Zendesk
 
@McNets ours is crap but works. I would have to look up the name but wouldn't recommend it anyway
For some customers we use theirs but I'm not impressed either by most of them
I'd say zendesk or Jira would be my pick
 
6:53 PM
This doesn't look like an answer. You can suggest it as an edit to the question (and delete this) — ypercubeᵀᴹ 23 secs ago
Does that make sense? (If yes, I can make the edit to the q.)
 
HI GUISE!
0
A: SSIS - Convert string variable to GUID

Hardik MasalawalaPlease try with Perfect Answer CONVERT(uniqueidentifier,'Your String Value') SELECT DocData FROM docsRepo WHERE DocumentGuid = CONVERT(uniqueidentifier,'53A394A7-5D2B-40C0-A04D-90553E4535C3')

"Perfect Answer"
 
good Lord, HELLLO BILL!
 
@ypercubeᵀᴹ hmm. It's between a comment an edit or even a new question
 
@billinkc perfect answer. nice.
 
:44217191 struggled with mobile. Sorry for the pingd
 
7:01 PM
@billinkc hmmm good point
something must be wrong with the continuum. None of the answers on that question have a downvote.
 
I don't know enough about the subject of scaling out but it happens we have the ssis guru in here so maybe bill can judge
 
@billinkc a wild Bill appears
4
 
@TomV Ssis guru? Oh, you mean Bill!
 
its funny how few people understand that the model database is used as a basis for each new database created.
 
7:39 PM
@billinkc you're still alive?
 
I suppose so
y tu?
 
7:54 PM
yo también, gracias
@billinkc I suppose you could be of some help
 
@MaxVernon like the template1 in Postgres.
 
8:40 PM
No woogy required
SELECT @x.query('/xml/customer[email/@address[not(contains(., "@"))]]') answer
2
 
oh who asked you, four eyes?
 
Getting woogy with it
Yes, yes I did that; dual Phil Collins AND Will Smith references in nearby comments. Believe.
 
i like it
 
 
1 hour later…
9:47 PM
@wBob ahh good old "not". Was wondering what the syntax for that was.
 

« first day (2668 days earlier)      last day (2190 days later) »