Re: Efficiently modeling sets and subsets in lattice-like structure
Hi Kelly,
I'm pretty sure there's no universal efficient solution to this
problem, this must be the well known answer
you are looking for. The pure mathematics of 'very hard' could be very
hard to express, though.
So before you choose a solution you need to find out more about the
actual graphs you're working
On the other hand there are plenty of possible approaches to the graph
storage problem.
They do seem to concentrate on trees (XML is one of the better ones).
So you could approach the graph representation
problem as a tree of trees, or as a set of separately stored trees
joined logically over common nodes.
I doubt you could find many solutions in that direction.
Then since a graph is a binary relation over a set of nodes you could
represent it as a 2 field relational table in an obvious way.
SQl isn't very good with trees, though.
In any case on model 'a Bunch Of Regular Guys' could code themselves (I
did it once!) is a representation
of the binary relation as a bit matrix. You'll need N*N bits where N is
the number of nodes so it's not going to scale well
but if you do happen to have fewer than 90K nodes (1G matrix) you could
use it directly. If you have more nodes but
the number of ones in the matrix is not too high you could compress the
matrix somewhat at the expense of extraction/update performance.
If you have many nodes with lots of connections you're out of luck
anyway.
For persistent storage in this case you could use a relational table in
few obvious ways, or just a file.
Example:
7 11 15 X Y 15 20 25 Z 14 30
1 2 3 4 5 6 7 8 9 10 11 12
maintaining this array is a separate issue (or non issue), here a two
field table in an RDBMS is definitely appropriate.
1 2 3 4 5 6 7 8 9 10 11 12
1 all zeroes
2 all zeroes
3 all zeroes
4 1 1 1 0 1 0 ... # 4 is X which is 7 11 15 Y
5 0 0 0 0 0 0 1 1 1 1 0 0 # 5 is Y which is 15 20 25 Z
6 and so on
7
8
9
10
11
12
now if your sets are ordered and {7,11} is not the same as {11,7} this
isn't going to work at all...
Thanks,
Michael
-Original Message-
From: Kelly Jones <[EMAIL PROTECTED]>Bunch Of Regular Guys
To: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED];
[EMAIL PROTECTED]; mysql@lists.mysql.com
Sent: Wed, 30 May 2007 2:53 pm
Subject: Efficiently modeling sets and subsets in lattice-like structure
Apologies for the mass cross-posting: I haven't been able to find a
single
answer or reference for the problem below (googling didn't help), and
was hoping someone could point me to something helpful. I'm convinced
there's a well-known answer here that I just can't find :(
We're modeling a collection of (finite mathematical) sets, where sets
may contain each other as subsets.
For example, set X may be defined as "{7,11,15} + Y" meaning X
contains 7, 11, 15, and all the members of set Y. Set Y could then be
"{15,20,25} + Z", and Z might be "{7,14,30}" with no subsets.
Of course, no set includes itself, directly or indirectly.
However, one set may include many other subsets (eg, X may include
both Y and Z), and one set may be included in many others (eg, X may
be included in sets V and W).
If the sets were nodes in a directed graph, the graph would be
acyclic, but not a tree. I believe this is called a "lattice"(?).
How can we efficiently model this "lattice" of sets either using SQL
or some other technique? Specifically:
% Add or remove members/subsets from a set and have the changes
"bubble up" the lattice efficiently.
% Find all members of a set X, efficiently traversing subsets. In
other words, find all nodes/leaves of the subtree rooted at X
% For a given set X, find all sets that contain X as a subset,
directly or indirectly (eg, if Y contains X, and Z contains Y, return
both Y and Z). In other words, find all the nodes that have X as a
sub-node.
The most promising lead I'd found was:
http://dev.mysql.com/tech-resources/articles/hierarchical-data.html
(which is why I'm cc'ing the MySQL list), but this only works with
trees, and I couldn't figure out how to modify it for lattices.
-- We're just a Bunch Of Regular Guys, a collective group that's trying
to understand and assimilate technology. We feel that resistance to
new ideas and technology is unwise and ultimately futile.
-- MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:
http://lists.mysql.com/[EMAIL PROTECTED]
Check Out the new free AIM(R) Mail -- 2 GB of storage and
industry-leading spam and email virus protection.
=0
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: spool data/log into a file.
most likely you just need to redirect STDERR to tee as well so it's not a mysql problem: mysql -u -p -f << eof 2>&1 | tee ttt.txt ... assuming it's Bourne or ksh, don't remember what's csh for 2>&1 Regards, Michael -Original Message- From: [EMAIL PROTECTED] To: mysql@lists.mysql.com Sent: Thu, 26 Apr 2007 8:15 AM Subject: spool data/log into a file. Hi All, I have written a script to connect to mysql and all the command results is spooled into a file. But when an error occurs in mysql, it stops at that point and does not proceed to the next command, also these error command is not written to the output file. Can you please tell me what i need to do, for the script to continue even if there is an error and that these errors should be spooled into a file. I have used -f(force) option while connecting to mysql db, it continues even if there is any error in the script, but it does not write error command into the output file. Can you please help how can i spool all the output into a file. Below is the simple script i am using, ttt.txt is the file which will contain the all the output of the script. mysql -u -p -f << eof | tee ttt.txt use dev1; select count(*) from rep_abc; select count(*) from rep_fact; exit eof Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Safe DB Distribution
I think in the software world pretty much anything can be done, not 100% to the specs but using the 'for all practical purposes' criteria. it's a matter of the cost but the PHP + mysql combination certainly can't protect the datamodel because most of it can be hacked from the data files themselves. So first thing the OP needs is a custom mysql code to make it difficult (not impossible). One thing that comes to mind is to disable all forms of caching to avoid getting much from the memory dumps. If he's willing to do this than switching from PHP to a binary executable will be a breeze, or the PHP script will use callouts to proprietary interface of the customized mysql. I think it would be safe to claim that all data in the world that is worth this kind of investment is already available either free or for a modest fee, think Linux or mysql source code or any good book you can by for a few dollars. Bigger products often have a funny and not at all intentional builtin mechanism against pirating where it's very hard to install and use the software unless you also buy a support contract. Thanks, Michael -Original Message- From: [EMAIL PROTECTED] To: mysql@lists.mysql.com Sent: Tue, 30 Jan 2007 1:52 PM Subject: Re: Safe DB Distribution [EMAIL PROTECTED] wrote in news:[EMAIL PROTECTED] R04.sysops.aol.com: Something like distributing files with encrypted data in most columns can be done I think the OP is more concerned with protecting his datamodel, and this cannot be done in my opinion. -- felix -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Safe DB Distribution
> You will not be able to search on encrypted fields. Right. There are ways around it but obviously all sorts of subtle application work will go with the data, depending on what exactly is that you are trying to protect. Overall my opinion is that if you distribute it in a country where software piracy is a norm, someone will hack it anyway, in all other countries it's probably less expensive overall just distribute it and hope that most customers will comply with the law. If you have competition you'd prefer people steal your stuff and not theirs. Thanks, Michael -Original Message- From: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED]; mysql@lists.mysql.com Sent: Tue, 30 Jan 2007 1:44 PM Subject: Re: Safe DB Distribution [EMAIL PROTECTED] wrote: Something like distributing files with encrypted data in most columns can be done, with the encryption key being a combination of some hardware-related stuff such as MAC address that the user will provide to you, and the key you generate and provide to the user. Of course it creates a problem when the target computer crashes. Everything else is probably either hackable or would be based on the assumption that you'll be in the business forever but anyway there must be specialists out there in some security related groups that know how to make it hard to hack. > On the surface of it if you provide this sort of protection you'll be > able to distribute it as open source since you are protecting data and not the code but I'm sure bigger lycensing minds will easily prove me wrong and explain that there is such a thing as a GPL for data and copyrighting data is bad, which would mean that copyrighting anything is bad, or that copyrighting data is OK but the essence of your application is data so if you charge for it you can't distribute mysql with it under GPL etc. Actually it would be interesting to hear the opinion of mysql legal > department on this. > Thanks, Michael -- > felix > -- > MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] Check Out the new free AIM(R) Mail -- 2 GB of storage and > industry-leading spam and email virus protection. > You will not be able to search on encrypted fields. -- Gerald L. Clark Supplier Systems Corporation Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Safe DB Distribution
Something like distributing files with encrypted data in most columns
can be done, with the encryption key being a combination
of some hardware-related stuff such as MAC address that the user will
provide to you, and the key you generate and provide to the user.
Of course it creates a problem when the target computer crashes.
Everything else is probably either hackable or would be based
on the assumption that you'll be in the business forever
but anyway there must be specialists out there in some security related
groups that know how to make it hard to hack.
On the surface of it if you provide this sort of protection you'll be
able to distribute
it as open source since you are protecting data and not the code but
I'm sure bigger lycensing minds will easily prove me wrong and explain
that there is such a thing as a GPL for data and copyrighting data is
bad,
which would mean that copyrighting anything is bad, or that copyrighting
data is OK but the essence of your application is data so if you charge
for it you can't distribute mysql with it under GPL etc.
Actually it would be interesting to hear the opinion of mysql legal
department on this.
Thanks,
Michael
-Original Message-
From: [EMAIL PROTECTED]
To: mysql@lists.mysql.com
Sent: Tue, 30 Jan 2007 4:08 AM
Subject: Re: Safe DB Distribution
[EMAIL PROTECTED] ("Suhas Pharkute") wrote in
news:[EMAIL PROTECTED]:
I want to distribute the MySQL DB to different user with an PHP web
app.
Make sure you have read and understood the MySQL licensing policy, if
your
app is not open source (as it seems to be).
My biggest problem is MySQL data files. Is there anyway by which
I can configure the MySQL so that only PHP script can read it and if
you happen to copy the Data files and try to use it, that will be
useless.
No.
--
felix
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
Check Out the new free AIM(R) Mail -- 2 GB of storage and
industry-leading spam and email virus protection.
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: General MySQL Question: Ed Reed (CA, United States of America) Medium
Hi Ed, the subject line of your message indicates you live in California in which case Santa Clara won't add much to the experience, but most other people could enjoy visiting California a lot especially if they don't mind dedicating entire day driving to Sierra Nevada. One off the beaten path route is to cross the Sierra on hwy 108, pass by Lake Mono and return on hwy 120 through Yosemite. Even around Santa Clara there are some possibilities for a good drive, try to get to the observatory in San Jose (hwy 103 or 104), or just cross the Coastal range around Santa Clara and take hwy 1 North to San Francisco. Other than the natural beauty the place is pretty sterile but from my experience it applies to most of the country anyway. The clubs in downtown San Jose are decent especially if you manage to find one with live music. I'd consider visiting one of mysql conferences if they had one in a nice place like Panama City (in Panama, not in Florida). Thanks, Michael -Original Message- From: [EMAIL PROTECTED] To: mysql@lists.mysql.com; [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Mon, 29 Jan 2007 4:57 PM Subject: Re: General MySQL Question: Ed Reed (CA, United States of America) Medium I just didn't enjoy the location. When it was in Orlando a few years ago, it was great. There were plenty of things to do and see; different places to eat every night. I had a really good time. Then the following year I went to Santa Clara and there was nothing to do. I went to the movies one night. Had dinner at Sizzler three times. I drove 45 minutes to find a place I could buy a shirt. The rest of the time I stayed in the hotel and watched crappy tv. If you folks want me to be away from my family and friends for five days you've gotta give me a reason to wanna go besides the great technical information. I didn't go to last years, I'm not going to this years and I probably won't go to any future ones held in Santa Clara. When I came home from the last one, my kids asked what I got them from my trip and I had nothing for them. It was just a boring place to go and I don't wanna go back. I'd just like to see it held someplace new every year. Thanks for replying to my comment Lenz Grimmer <[EMAIL PROTECTED]> 1/25/07 2:21 AM >>> -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi Ed, thank you for your message! On Thu, 18 Jan 2007, Ed Reed has filled out the MySQL Contact Form with a General MySQL Question: Can someone tell me you folks are going to hold all future User Conferences in Santa Clara? I would like to go this year but I didn't enjoy myself at the last one I went to in Santa Clara in '05 and I don't wanna go back. If you folks plan to hold all future conferences there I'd like to know so I can resign myself to the fact that it's never gonna change instead of hoping that it's gonna be somewhere else every year. To be honest, it's not entirely clear yet if we will change the location for upcoming events. But could you explain why you didn't enjoy yourself at the last one? How can we improve? Bye, LenZ - -- Lenz Grimmer <[EMAIL PROTECTED]> Community Relations Manager, EMEA MySQL GmbH, http://www.mysql.de/, Hamburg, Germany Visit the MySQL Forge at http://forge.mysql.com/ -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (GNU/Linux) Comment: For info see http://quantumlab.net/pine_privacy_guard/ iD8DBQFFuISeSVDhKrJykfIRAsJxAJ9hLkjhA10B68OHpw97x6QHbSqjLACePQym m9lEE93nxdC+K//cQsL658Q= =W/8N -END PGP SIGNATURE- Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: [OT} How to pronounce GIF
Hi, gif will inevitably gravitate towards jif for the same reason SQL is sequel and lib, as in /usr/lib, is often 'libe'. The reasons are described in the Appendix to Orwell's 1984 which is dedicated to newspeak, look for 'duckspeak' in particular, and the consideration of 'Minitruethful' vs 'Minitrueful'. In other words the pronunciation will be such that chances of the other party to actually think about its meaning be as little as possible. This is very good in a job interview situation (I have no idea what TCP/IP means but I desperately need this job...) or in IT salesman pitches which most of the time amount to 'give us your money in exchange for a lot of frustration trying to use our product' but you're far more likely to hear words like Java, Internet or even whole nice sounding expressions like 'closing the digital gap' or 'business intelligence' or 'globalization'. Giraffe, on the other hand, could be pronounced in every which way possible unless uttering it can gain you an advantage, I guess if you are biologists discussing giraffe breeding. Thanks, Michael -Original Message- From: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: mysql@lists.mysql.com Sent: Fri, 12 Jan 2007 2:04 PM Subject: Re: [OT} How to pronounce GIF Brian Dunning wrote: On Jan 7, 2007, at 4:23 PM, TK wrote: >> In short, the original inventors of the GIF format (CompuServe, 1987) >> have always defined the pronunciation to be like "JIF." So, that has >> always been the "correct" pronunciation. > > Sure, so I'll start pronouncing "graphics" as "jraphics". > How do you pronounce "giraffe"? -- Gerald L. Clark Supplier Systems Corporation -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: mysql on a NAS
Hi, OCFS is a cluster filesystem so running two mysqld over the same OCFS directory is exactly the same as running two mysqlds over the same directory in a local filesystem on the same machine which is strictly prohibited. If you decide to consider mysql clustering over ocfs please keep in mind thad you don't really need cluster filesystem for mysql clustering. mysql clustering doesn't share disk storage. Considering they use ocfs they may also be mistaken about the difference between mysql and oracle clustering. Oracle cluster requires shared storage whereas mysql does not. Thanks, Michael -Original Message- From: [EMAIL PROTECTED] To: mysql@lists.mysql.com Sent: Tue, 28 Nov 2006 12:59 PM Subject: Re: mysql on a NAS On Tue, 2006-11-28 at 17:58 +0100, Stefan Onken wrote: Can you explain this a little bit more ? I am not the guy who set it up, so I would like to go back them and say "Well, You cannot do this, because... " :) http://dev.mysql.com/doc/refman/5.1/en/multiple-servers.html describes the pitfalls of running multiple servers on the same machine about halfway down the page in the bold 'Warning' paragraph. -- Pat Adams Digital Darkness Promotions Dallas Music Wiki http://digitaldarkness.com -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: finding NULL records
Thanks, I just couldn't remember English for 'ternary'. Remembering the and/or tables for the logic became trivial when I replaced 'NULL' for 'UNKNOWN' but 'UNKNOWN' doesn't exist as an SQL keyword. I suspect the founding fathers chose NULL over words like UNKNOWN to try avoid philosophical debates since there are at least two meanings for null values, one that the data exists but is unknown, and another that the data simply doesn't exist, such as date of death for a live person. But I think this doesn't really belong to a mysql forum but rather to a general SQL discussion so I'll try not to comment on that anymore. Thanks, Michael -Original Message- From: [EMAIL PROTECTED] To: mysql@lists.mysql.com Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED] Sent: Thu, 16 Nov 2006 2:55 PM Subject: Re: finding NULL records Hi all, may I propose some slight corrections in wording ? Inserted below at the appropriate places: Jerry Schwartz wrote (re-ordered): >> -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Wednesday, November 15, 2006 2:22 PM To: [EMAIL PROTECTED]; mysql@lists.mysql.com Subject: Re: finding NULL records [[...]] This is what happens when one of the most innovative concepts of SQL gets overlooked or understressed which it seems to be in most modern SQL books. The thing is that SQL's logic isn't binary. Rather it's built over three values: TRUE, FALSE, NULL. Nearly: The three values are typically called TRUE, FALSE, and UNKNOWN. A logic based on three values is then called "ternary", AFAIR. You can indeed compare anything to NULL: You sure can, like you can compare apples to oranges - just the result is not meaningful. In SQL, NULL is no value, it rather describes the absence of any value. What often gets overlooked: "absence of value" is different from "empty", so an empty string is a non-NULL value ! x = NULL in most systems is a valid syntax. The result is NULL if x is UNKNOWN [[...]] Even if it is valid syntax, you cannot really use it: Comparing something known (some value) to something unknown (no value present) cannot give you any definite result, so in SQL for *any* comparison operator "op" ( = < > <> ... LIKE ... ) and *any* known (= non-NULL) value x (column, literal, expression, ...) the result of x op NULL is UNKNOWN. As a rough analogy, consider comparing some visible object to one which is hidden behind a curtain: You cannot tell whether they are identical or not, which one is larger, ... , so the result is UNKNOWN. Especially, the result is *not* FALSE ! If such a predicate is used in a WHERE condition, like in SELECT x, y, z FROM tab WHERE x op NULL then "x op NULL" evaluates to UNKNOWN, this is different from TRUE, and so the WHERE condition is not satisfied. x IS NULL on the other hand isn't a binary operator, it's an UNARY one with the result values of TRUE or FALSE. Correct. The whole join theory obviously revolves around NULL treatment but otherwise the NULL algebra becomes fuzzy. I'm not sure what the standards say to the question of sorting NULLs, or whether each null constitutes a separate group as in GROUP BY, or all can be grouped together, but the actual implementations often make the answers configurable. I don't have the standard at hand either, the only thing I remember is that it requires NULL to sort either less than or greater than all non-NULL values. Beyond that, it often becomes implementation-dependant. [[...]] The result of all but two expressions will be NULL (not 0 or 1) if either UNKNOWN side is NULL. The two exceptions are the operators IS NULL and IS NOT NULL. > SELECT NULL = NULL; /* result is NULL */ UNKNOWN SELECT NULL IS NULL; /* result is 1, or TRUE */ SELECT NULL IS NOT NULL; /* result is 0, or FALSE */ Regards, Joerg -- Joerg Bruehe, Senior Production Engineer MySQL AB, www.mysql.com Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: finding NULL records
Well, I'm not sure about the exact definition of 'all'. NULL OR TRUE /* result is TRUE, example (NULL = NULL) OR (1 = 1) */ and so on. Didn't try to execute the actual SQL with such a close, though. Thanks, Michael -Original Message- From: [EMAIL PROTECTED] To: [EMAIL PROTECTED]; [EMAIL PROTECTED]; mysql@lists.mysql.com Sent: Wed, 15 Nov 2006 3:04 PM Subject: RE: finding NULL records The result of all but two expressions will be NULL (not 0 or 1) if either side is NULL. The two exceptions are the operators IS NULL and IS NOT NULL. SELECT NULL = NULL; /* result is NULL */ SELECT NULL IS NULL; /* result is 1, or TRUE */ SELECT NULL IS NOT NULL; /* result is 0, or FALSE */ Regards, Jerry Schwartz Global Information Incorporated 195 Farmington Ave. Farmington, CT 06032 860.674.8796 / FAX: 860.674.8341 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Wednesday, November 15, 2006 2:22 PM To: [EMAIL PROTECTED]; mysql@lists.mysql.com Subject: Re: finding NULL records trying to revive an old thread for fun... > I inherited a system in which the code was writing the string "NULL" into a > CHAR field. This is what happens when one of the most innovative concepts of SQL gets overlooked or understressed which it seems to be in most modern SQL books. The thing is that SQL's logic isn't binary. Rather it's built over three values: TRUE, FALSE, NULL. You can indeed compare anything to NULL: x = NULL in most systems is a valid syntax. The result is NULL if x is not NULL, and the result of NULL = NULL I honestly don't remember, but it's either NULL or FALSE, definitely TRUE. The same goes for any other binary operator I can imagine. x IS NULL on the other hand isn't a binary operator, it's an UNARY one with the result values of TRUE or FALSE. The whole join theory obviously revolves around NULL treatment but otherwise the NULL algebra becomes fuzzy. I'm not sure what the standards say to the question of sorting NULLs, or whether each null constitutes a separate group as in GROUP BY, or all can be grouped together, but the actual implementations often make the answers configurable. I wish SQL educators paid a little more attention to the math of it. On the other hand describing joins in a kind of 'traditional' mathematics [no pictures, just symbols] is definitely an overkill. Thanks, Michael It took me a long time to figure out what MySQL CLI was showing > me. > > These kids today... Oh my, sounds like a candidate for www.thedailywtf.com :-) Martijn Tonies Database Workbench - development tool for MySQL, and more! Upscene Productions http://www.upscene.com My thoughts: http://blog.upscene.com/martijn/ Database development questions? Check the forum! http://www.databasedevelopmentforum.com -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] __ __ Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: finding NULL records
trying to revive an old thread for fun... I inherited a system in which the code was writing the string "NULL" into a CHAR field. This is what happens when one of the most innovative concepts of SQL gets overlooked or understressed which it seems to be in most modern SQL books. The thing is that SQL's logic isn't binary. Rather it's built over three values: TRUE, FALSE, NULL. You can indeed compare anything to NULL: x = NULL in most systems is a valid syntax. The result is NULL if x is not NULL, and the result of NULL = NULL I honestly don't remember, but it's either NULL or FALSE, definitely TRUE. The same goes for any other binary operator I can imagine. x IS NULL on the other hand isn't a binary operator, it's an UNARY one with the result values of TRUE or FALSE. The whole join theory obviously revolves around NULL treatment but otherwise the NULL algebra becomes fuzzy. I'm not sure what the standards say to the question of sorting NULLs, or whether each null constitutes a separate group as in GROUP BY, or all can be grouped together, but the actual implementations often make the answers configurable. I wish SQL educators paid a little more attention to the math of it. On the other hand describing joins in a kind of 'traditional' mathematics [no pictures, just symbols] is definitely an overkill. Thanks, Michael It took me a long time to figure out what MySQL CLI was showing me. These kids today... Oh my, sounds like a candidate for www.thedailywtf.com :-) Martijn Tonies Database Workbench - development tool for MySQL, and more! Upscene Productions http://www.upscene.com My thoughts: http://blog.upscene.com/martijn/ Database development questions? Check the forum! http://www.databasedevelopmentforum.com -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Group By question
select * from t where emailaddress in (select emailaddress from t group by emailaddress having count(*) > 1) order by emailaddress; Thanks, Michael -Original Message- From: [EMAIL PROTECTED] To: mysql@lists.mysql.com Sent: Wed, 30 Aug 2006 5:17 PM Subject: Group By question I have a table of people with one of the fields being the email address. I would like to query all rows that have more than one person with the same email address. For example if the data were like this... A [EMAIL PROTECTED] B [EMAIL PROTECTED] C [EMAIL PROTECTED] D [EMAIL PROTECTED] E [EMAIL PROTECTED] F [EMAIL PROTECTED] The query would return row A, D, B, and E, in that order. It would not return C or F -- Chris W KE5GIX Gift Giving Made Easy Get the gifts you want & give the gifts they want One stop wish list for any gift, from anywhere, for any occasion! http://thewishzone.com -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: mysqlmanager logging?
Hi, you sort of suggest to look further beyond file permission problems but if I may: > drwx--x--x 2 mysql mysql 1752 2006-05-01 09:33 mysql doesn't look quite right, lack of read permissions to group/other. Thanks, Michael Izioumtchenko -Original Message- From: sheeri kritzer <[EMAIL PROTECTED]> To: Duzenbury, Rich <[EMAIL PROTECTED]> Cc: mysql@lists.mysql.com Sent: Fri, 5 May 2006 16:00:05 -0400 Subject: Re: mysqlmanager logging? su - mysql touch /var/lib/mysql/mysqlmanager.log see if that helps; maybe having the file there will kick it into gear. Is mysqlmanager actually running? Is there a pid file? -Sheeri On 5/5/06, Duzenbury, Rich <[EMAIL PROTECTED]> wrote: > > > > -Original Message- > > From: sheeri kritzer [mailto:[EMAIL PROTECTED] > > Sent: Thursday, May 04, 2006 3:12 PM > > To: Duzenbury, Rich > > Cc: mysql@lists.mysql.com > > Subject: Re: mysqlmanager logging? > > > > Can the program write to /var/lib/mysql/mysqlmanager.log? > > check permissions. > > > > # su - mysql > [EMAIL PROTECTED]:~> whoami > mysql > [EMAIL PROTECTED]:~> cd /var/lib/mysql > [EMAIL PROTECTED]:~> touch foo.txt > [EMAIL PROTECTED]:~> ls -al > total 3 > drwxr-xr-x 4 mysql mysql 120 2006-05-05 13:25 . > drwxr-xr-x 56 root root 1472 2006-05-05 04:25 .. > -rw-r--r-- 1 mysql mysql 0 2006-05-05 13:25 foo.txt > drwx--x--x 2 mysql mysql 1752 2006-05-01 09:33 mysql > drwxr-xr-x 2 mysql mysql 48 2006-05-01 09:33 test > > I don't think there is a permission problem. > > Any further ideas? > > Thanks. > > Regards, > Rich > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] ___ Try the New Netscape Mail Today! Virtually Spam-Free | More Storage | Import Your Contact List http://mail.netscape.com -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
