@billinkc The usual <verb> <noun> naming convention suggests your package "mvp_PeterLarsson.dtsx" takes well-known Swede Peter Larsson and converts him into an MVP (probably via a Script Task I would guess)?
@wBob It's segmentation by source material - in my SO project, I have packages that take the form of so_QuestionID, dba_QuestionID and for the rare question stemming from the MVP email group, I just use mvp_Questioner
@billinkc lol I was making a kind of rubbish SSIS joke, I guess it doesn't work (like my SSIS error handling framework : )
@billinkc SwePeso is of course famous, although I guess he doesn't hang around here. Ahead of his time I think, that kind of hard technical content wasn't as popular back then, although he seems to have gone quiet these days.
I got the joke and smirked. He was having an interesting yet sadly unreproducible error, for me, with unicode and SSIS
The first Swede the comes to mind for me would probably be Erland but that could be based on how opinionated he seems to be, when it comes to SQL Server
i have a developer who's seeking on a range within our partitioning key ... partition key is datetime ... developer is casting as date .... he's scanning all partitions with nasty lookups
cast('' as datetime) encourages partition elimination
(the determinism thing was just a rabbit hole i fell through while trying to answer the developer's question ... "isnt date just a subset of datetime?")
I caught my trainee copy pasting code from that site a while back
I told him "it's ok if you know what the code is doing but forgot the syntax, it's not ok if you just copy paste that without any explanation about how and why it works and what the gotcha's are"
because that's the main issue in my opinion, no context, no gotchas, just a bit of code
Given the pagerank I'm sure a lot of that is in the wild
On a brighter note, I started using a new strategy to analyse blocking and make it visual for the business, it seems to work well and impress so-called managers
shredding the blocked process reports XML, copy paste into excel and create nice pivotcharts
seems to work well to explain to the business why it's not necessarily a technical issue, also it gives a quick overview about the domains where the blocking occurs
@wBob No data type, you mean? The rest seems pretty good for v1. More than I expected actually.
That's insane. I don't mean your solution - very good - but it's insane that we need to do ridiculous things like this just to drop an index! — PhilHibbs3 hours ago
Converting a process that uses 3 stored procs with RBAR which runs for over 4 minutes down to 500ms in a single SQL statement with one index, on SQL Server 2000, is what I call winning today. The business rejects it because it will take time to test. Sad trombone.
I especially love that it has a function to strip the time off a date, and the date happens to be today, over and over and over. That doesn't make it slow at all, nope.
I developed an SSIS package about 6 months ago to migrate a number of access databases into SQL server. I opened the package yesterday to go through another run and noticed a few (X) error indicators. Upon further investigation, the connection managers produced the following error:
"The specif...
Any of you folks know of any good articles comparing performance of Full/Simple/Bulk recovery models.
I've got some tests that I've done but I'd like to offer as much evidence as possible that we have no business keeping an overnight loaded Data warehouse on Full Recovery.
@MasterDatabase They seems to had added this to the wrong place.
@JamesAnderson Good point. For that purpose, I often put a --SELECT * on a separate line in an UPDATE statement just before the FROM part, so I can select the part starting from the SELECT till the end of the statement.
You don't really need to uncomment it, it can remain in the UPDATE for future debugging since it's just a comment. But you do need to add the line itself.
I mean something like this:
UPDATE
tgt
SET
tgt.col = src.col
--SELECT *
FROM
tgt
INNER JOIN src ON src.keycol = tgt.keycol
WHERE
...
;
@ypercubeᵀᴹ Do any of the other RDBMSs throw an error when there are multiple source rows for a target row? the MERGE in T-SQL does but not the UPDATE with JOIN.
There is an 8 character limit on some things in oracle, try shorting database1, database2 and database3 to db1, db2 and db3 also the characters _ and # are allowed if i remember correctly.
all other things equal, and in the interest of fun and science, you need to return 168 rows from a database and then send them to the client. Do you use a dataset or a datareader, or who cares, for the task?
Is there anything fundamentally wrong with this postgres query?
BEGIN UPDATE "Predictions" SET "Joker_Played" = False WHERE user_identity = "Predictions"."userid"; UPDATE "Predictions" SET "Joker_Played" = True WHERE "Predictions"."Fixture_No" = "Web_Fixture_No" AND "Predictions"."userid" = user_identity; END;
@JamesAnderson I remember discussing - here in the chat - this ill behaviour of UPDATE in SQL Server. It may even update the same row multiple times. Not sure what other DBMS do in that case. MySQL probably allows it too.
I'm not asking for a particular reason; I have a decent I think understanding of the constructs. I chose a reader because there was no processing (we're just adding results to a list). Someone else changed it to a dataset despite having had ample opportunity to ask me to change it over the past few days. Mildly annoyed, because in so doing said person stepped on several changesets worth of stuff I had added to the method.
I mean I think you argue it either way but in this specific case I don't think it's a "you must use one".
@ypercubeᵀᴹ I need the two where statements to remain different. the first update statement is supposed to set all rows for that user to false. Then the second update statement is supposed to set specific rows to true
@ypercubeᵀᴹ It actually uses an any aggregate to just pick a value to use. Seems a bit odd to me. At least that means it doesn't waste writes by doing multiple updates to the same row.
The end result would be equivalent but the rows would be updated once (while with 2 statements, some rows would be updated twice, and if there are triggers, the side effects might be different)
@LemusThelroy It wouldn't. It would set the values directly to the wanted one (either True or False)
so I have a tiny brain teaser. If you have two SQL Servers, one prod and one non-prod, and the non-prod server is returning data 3 times quicker than the prod box, how do you diagnose the problem. Oh, did I mention, this is SQL Server 2000 - 'cuz I know how to do it on 2005+ :-)
I know about sysprocesses but not much else in terms of system views on 2000.
I am using postgres 9.4.6 and very new to postgres and any form of sql.
I want to create a query that updates all of the values for a user to False and then runs a query that updates specific values to True.
I have creating the following statement which is not updating values to True which has ...
Today I saw these two questions. They are IMHO very low quality/effort ones, but may be "fixed" with a few comments and right response from OPs. But looking at them together, they are two variants of the same thing, with the same wrong assumptions yet comming from two different (new) users. If it...
Looks like the Oracle answerer at least has a reference for what they were talking about. Now, whether that ties in with the original question is beyond me
this may sound silly, but is there any postgres communities where you can be mentored by someone / chat to someone one-to-one rather than posting a million quetsions or risking getting humiliated in a chat room? lol
I would look at places like meetup and possibly whether PG has chapters like the SQL Server community does. If not, maybe it's your calling to start one in your community.
I studied python for the last 9 months and decided to write a web app, which requires some basic postgres. However, I have to get the application done by 10th May and my basic postgres queries are falling on their face!
I wish I knew that I couldn't have got away with taking time to learn postgres properly
The whole reason there is a SQL Saturday event in Kansas City, where I live, is because I wanted training and couldn't find what I wanted. Start small, ask around at staffing firms if they have people that do PG work and if they'd be interested in sponsoring room/pizza for like minded people to get together and chat
Seems like a long time to be learning python - your first language?
we've had a similar definition: done (when devs say it's done), done done (when it's actually checked in), and done done done (when a tester validated it) :)
Not knowing and not wanting to know the specifics of your website, I would think there'd be a scaffolding/framework for Create, Read, Update, Delete (CRUD) operations that you could leverage in helping you through the database procedure portion
You know that "Slow Delete's of LOB data in SQL Server" question, would it not be worth recommending partitioning, with partition switch and truncate? I know they haven't posted edition, and I know partitioning has its ups and downs but could it be an answer?