@Phil mine were requisitioned as soon as they arrived - I just ended up with the shirt - but then I needed new pyjamas :-)
.
for those who care about such things, we are getting a fair bit of traffic and it's been growing steadily for quite a while. We seem to have got over our dependence on SO migrations (that stopped abruptly earlier this year).
when i looked at this yesterday it was the postgres part that was a pain. looks like it's easy to export to base64, but a nightmare for straight binary
Is This something you are looking for ???
Create Trigger RecordChange on P1
after insert, update
As
begin
declare c_get_row cursor
for
select MsgNumber,CardNumber,Action from inserted
declare @MsgNumber int,@CardNumber int,@Action varchar(10)
open c_get_row
fet...
@MarkStorey-Smith What Lamak means (and I agree), is that this IN list, `IN ( CASE @IsPaid WHEN 1 THEN 1 WHEN 0 THEN 0 END, CASE WHEN @IsPaid IS NULL THEN 0 END, CASE WHEN @IsPaid IS NULL THEN 1 END )`, evaluates to `IN (NULL, 0, 1)` when `@IsPaid` is NULL.
This issue came up when I got different records counts for what I thought were identical queries one using a not in where constraint and the other a left join. The table in the not in constraint had one null value (bad data) which caused that query to return a count of 0 records. I sort of unders...
I'm reading the book "SQL Server 2008 Query Performance Tuning Distilled" and found its description of perf mon metrics to be a little vague in some cases. I'll list all the metrics it covered, with my sense of each. I'd like to have clarifications shared where necessary. As my comments sugges...
Wow, I don't think I've ever seen ALL() syntax used in the practical world before. But I find the logic significantly complicated to parse. Anyone else?
@Phil this is actually one of the typical causes for "slow in the application, fast in SSMS" - since the procedure is usually created in SSMS, the creation-time and runtime settings match. C# uses different default settings, no match = recompile or poor plan choice
@Lamak it's terribly vague terminology and look at the clause that's require to meet eligibility. What does WHERE 1 = ALL (SELECT ... FROM ... WHERE something <> 0) mean? Can you express that easily in English? I can't.
@swasheck not really, optimization-wise. Mostly if you write a function as 'LANGUAGE SQL STABLE' then it gets inlined anyways and all the same stuff happens
@rfusca I still find EXISTS cleaner and more self-documenting, and it's not as intuitive to figure out what's going on without his additional comment, but I agree I'm actually impressed that someone was able to find a practical use.
@AaronBertrand Ya, I think EXISTS is more 'standard' but I thought the use was fairly obvious once I saw it. Otherwise the logic becomes 'NOT EXISTS (SELECT --the part not resolved)' and I find that more convoluted.
TNSLSNR for IBM/AIX RISC System/6000: Version 10.2.0.1.0 - Production System parameter file is /oracle/10.2.4/db/OraHome_1/network/admin/listener.ora Log messages written to /oracle/10.2.4/db/OraHome_1/network/log/listener.log Listening on: (DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=EXTPROC))) Error listening on: (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=testiold)(PORT=1521))) TNS-12542: TNS:address already in use TNS-12560: TNS:protocol adapter error
Am I worrying pointlessly here? Somehow I feel uneasy about making any assumptions about inexact types:
I wonder if at least in theory it is possible that random() would return a value <1 that when multiplied by 9 would be >=9 due to the inexact nature of the double precision type? In practice even if it is possible it would be vanishingly unlikely of course because of the 15 digits or so precision. — Jack Douglas11 mins ago
@JackDouglas I don't think this is possible with multiplication by a whole number (which is exact), because the number of total decimal places cannot increase.
@JackDouglas .2 is inexact in IEEE fp because it is 1/5 and 5 is not a power of 2. So the binary representation will be non-terminating. So you are saying it's really .20000000000000000001 or whatever and then when multiplied by 9 it/s really 1.80000000000000000009
@JackDouglas So you're worried about getting .9999999999999999999 (whatever binary equivalent - by this point exact/inexact conversion to/from decimal is not an issue because the random is already an fp binary entity) and then multiplying that by 9 and getting something > 9?
I realise this may be come a subjective mess, but I'm hoping for some objective information on the subject. I also don't want to go shopping, I just want to know 'what to shop for'.
I'm very new to database management. As in I started SQL 4 weeks ago, and using MS Access last week. I'm coming to...
@Phil i love this "As in I started SQL 4 weeks ago, and using MS Access last week. I'm coming to the end of out perrenial database project using MS Access wherein I've hit up against a lot of the limiations to Access, and I've mentioned that we need something different."
@JackDouglas I'm not 100% sure, no. en.wikipedia.org/wiki/… I guess this would probably be a good question for the Computer Science SE. Basically can an fp <1 (I believe this can be known exactly) be multiplied by some integer in fp form to get a number larger than that integer due to rounding?
I would like to use code that I developed in C Sharp CLR to be used in all of the databases on the system so that I don't have to set each to trustworthy and turn CLR on and keep a bunch of the same code inside each one.
Is there a best way to do this from an administrative and security standpoi...
@JackDouglas It's databasey. We implemented plenty of CLR at my last job and we only ever implemented one central copy of the CLR assemblies. You can create wrappers that EXECUTE AS to avoid the trustworthy bit, but personally I don't understand all of the objection to trustworthy in the first place. And we had so much cross-db crap going on (mostly service broker) that we just turned on trustworthy anyway.
As for the backup / restore, the DLL is stored in the file system, not the database, so yes the DLL has to move. If the file path differs, then it needs to be updated, though I've never done that so I'm not sure how to re-register an assembly only updating the path.
Unfortunately I don't have time to craft a proper response now (headed to a Dr's appt) but if anybody wants to take my text here and form some kind of answer or comment, I won't object.
@JackDouglas Is this a topic y'all are particularly interested in? Sounds like he is primarily interested in the logistics of maintaining such a library, not so much actually writing it, which would point in your direction.
It seemed like much of the logistics he's interested in are how to implement the CLR library from a security perspective inside of SQL Server.
e.g. how to let databases call a central CLR assembly without turning on the trustworthy bit. That seems to have very little at all to do with source control.
so if i understand the question ... he wants to include CLR ... but only have [dbo] be able to execute (from any db) ... i may be naive but is that just DENY ... GRANT?
@JackDouglas so I need to be able to send databack secure to an app server; I need to secure all external connections except for one(appserver) and encrypt the data
@swasheck I know how to setup and use it
create users
tables etc
and dbs
but I need to learn cover my bases from making noob mistakes
The following configuration works for me on Oracle 11.2.0.2:
C prog:
[oracle@node1 ~]$ cat exec_hmmftg.c
#include <stdio.h>
int exec_hmmftg ( char *p1 , char *p2, char *p3 )
{
return 5;
}
[oracle@node1 ~]$
Compile it, and make sure it's placed in $ORACLE_HOME/bin or $ORACLE_HOME/lib ...
What are the main differences between MSSQl and MYSQL, other than the case of storage engines in mysql.??
Is This something you are looking for ???
Create Trigger RecordChange on P1
after insert, update
As
begin
declare c_get_row cursor
for
select MsgNumber,CardNumber,Action from inserted
declare @MsgNumber int,@CardNumber int,@Action varchar(10)
open c_get_row
fet...
I'm reading the book "SQL Server 2008 Query Performance Tuning Distilled" and found its description of perf mon metrics to be a little vague in some cases. I'll list all the metrics it covered, with my sense of each. I'd like to have clarifications shared where necessary. As my comments sugges...
You can try this demo:
CREATE TABLE #ATMStatusHistory (Code INT, RawFeed
INT, Resolved INT, Ticket INT)
INSERT #ATMStatusHistory VALUES
(2341283 ,2350426 ,0 ,205656),
(2341460 ,2350603 ,1 ,205656),
(2341597 ,2350740 ,1 ,205656),
(2341689 ,2350832 ,1 ...
Declae @sum int
SELECT @sum=SUM(value) FROM @t WHERE condition = 1
select *,@sum from yourtable
The following configuration works for me on Oracle 11.2.0.2:
C prog:
[oracle@node1 ~]$ cat exec_hmmftg.c
#include <stdio.h>
int exec_hmmftg ( char *p1 , char *p2, char *p3 )
{
return 5;
}
[oracle@node1 ~]$
Compile it, and make sure it's placed in $ORACLE_HOME/bin or $ORACLE_HOME/lib ...
