Can this be tweaked in a better way??
hi, ive picked up most things about OC. i built this pc and after a couple of months ive seen that the cpu is letting down my gpu HARD!!. so instead of forking out an extra 500bucks for 500 mhz i OC, i started this morning and i managed after numerous CMOS resets, i got my 3000+ (1.8ghz) cpu up to 2.3ghz equivilent to 4000+, im so happy.. i will list my rig below and what i have done in the bios. it seems pretty stable, ive played a couple games to test it out.. am i doing anything wrong that could harm the parts.. thanks, any better recommendations please.. I think my ram (216mhz +) is limiting my fsb, once i hit 260 its unstable at times, very weird...
MSI k8n neo4 FI - 1.60V
Amd 64 3000+ venice running at 2.3ghz (stock 1.8ghz) -FSB-255, (STOCK 200), Multiplier x8.5, cpu over vid 6.6%, HT x3, CPU VID 1.450V,
2X512MB Corsair xms 3200cl2 DDR400- auto settings, MEM 2.85V --Dont undestand how to change mem settings like 2-3-4-5 format.. MEM is running at 216mhz
XfX 7800GTX factory oc 490/1300
74gb Raptor 10,000 rpm
500wt Antec (ACM) - dont really know much about the voltage things.
2X120mm fans and 90mm fan.
3dmark 05 stock 7100
3dmark 05 OC 7850
What are your temperatures? With the CPU voltage and Mem voltage High, you have to be worried about exceeding maximum temperatures for the components. Do you have heat spreaders on the ram? or a aftermarket cooler on the CPU?
I try to keep temperatures below 50c for the board and CPU, they can run hotter, but not always the best Idea.
Im not experienced with your CPU, but I dont think your memory is causing the instability at 260 FSB. Your probably reaching the maximum overclock on your CPU. Try droping the multiplier on the chip, the multiplier on the HT down to 3x, and setting the memory to a 1:1 divider and seeing how far your FSB will go. If it stops at FSB 260 (Mem 216) then I would say yes you need to change memory timings or divider. But my guess would be that you can get that memory up to 240 or so before you see problems in MemCheck.
Refer back to the Overclocking guide for other tips.
Similar Messages
-
Can some one tell me a better way...
Hi All..
I work on a web application. I create user ids for people logging into my application. here I have a small problem.. This is what I am currently doing...
When I create a new user, I assign a new user id to the user and store all hi info.
All our user ids are stored in User_ID table and user info in User_Info table.
First I get the max user id from the User_Id table.
int iuserid = select max(User_ID) from User_ID;
Then I add ' 1 ' to this and assign it to new user.
iuserid = iuserid+1;
insert into User_id values(iuserid, <ssn> );
Then I store all user info in User_info table..
insert into User_info(iuserid, <first_name>, <last_name>,...);
Both these SQLs are executed as a transaction.
The problem that I have...
It works fine in a normal environment and in my testing.
But when the load inceases, before I insert into the User_info, another user tries to get the max User_id from User_ID table, then it conflicts..
I have seen occurences of User_Info storing the info of a different user against the User_id when many people are accessing the app at the same time.
Can some one tell me a better way to handle this scenario..
Appreciate all you help..
TIA..
CKHi,
assuming that the requirement for user_id is only for uniqueness (primary key requirement) not continuosly.
perhaps can try this,
1) create a table to keep the max id for each table's key.
e.g.
create table key_id_lookup
key_name char(n),
current_value number(m)
where n & m is the size of the field
2) for each creation of entry in the user_id table, lookup the value in the key_id_table and increment it after retrieval.
e.g. current_id = select current_value from key_id_lookup where key_name = 'user_id_key';
current_id+=1;
update key_id_lookup set current_value = current_id where key_name = 'user_id_key';
something similar to use of sequence if your database support it.
3) continue with the creation of record in other tables. now obtain the time stamp of creation and append with the current_id obtained. e.g. timestamp = 132456872; size of m is 5, user_id = 132456872 x 100000 + current_id;
this should give you a unique key at creation time.
There should be other better ways of resolving this(like locking the table for the update operation but affect performance, etc), depending on the feature supported by the database in use.
Hope this can help as last resort if other options not available. -
(CSS) Can this simple layout idea be better coded?
Hi guys,
I'm trying to build a very simple layout made slightly more complex via a background gradient that needs to be in sync with the foreground image.
http://vilverset.com/murray.php
As you can see, the layout is supposing a 1000px wide viewport, or bigger. The foreground image is 650px wide, and the text column next to it is 350px wide. This fills up our 1000px viewport.
Now, normally, I would just apply the image as a background to the text column's container (and make that container 1000px wide), except that there are a few conditions to this layout that make that solution ill-advised.
When there's enough content to fill it, the text column needs to stretch from the top to the bottom of the viewport, and trigger the main browser vertical scrollbar when exceeding this window.
When there isn't enough content to fill the length of the column, the content should be vertically centered.
The image on the left of the column should be vertically centered at all times, or it will lose its sync with the site's background gradient.
The image on the left of the column should never scroll with the content. It should be fixated.
The combination of the image and column should be centered in the viewport (both axis).
You've seen how I did it @ the link above. Div class "image" has the main image as a background, and this image is padded to extend to fill the whole 1000px width of the site layout. It's the only way I could think of to offset the image from the center spot. Ditto the text column, which is in a 350px wide div, that has a 650px left side margin (totalling 1000px).
The main negative side-effect of this method occurs when you reduce the viewport to less that 1000px in width. The image starts moving AWAY from the column (as you can see by trying it out at the link above).
If there's a better way to get that image and that column into position, and the combination of both centered on the screen in both axis... I'm willing to try it!I'm glad you're pleased with your layout. I'm having a few problems with it.
When viewed on a narrow screen such as an iPhone in landscape mode, your image moves leftwards out of view. And when device is scrolled horizontally to see content, text collides with the image. Is this what you want to have happen? (see screenshot)
Nancy O.
Alt-Web Design & Publishing
Web | Graphics | Print | Media Specialists
http://alt-web.com/
http://twitter.com/altweb
mjyeager wrote:
You tell me these things can't be done, but I've got foreground JPGs with gradients perfectly in sync with the background gradient, and neither of them scrolls with the text content. So not only can it be done, I've done it. Right there, at that link I posted in the OP. And it works in all major browsers, including Safari on Mac.
I was just wondering if there was a BETTER way to do it. But if people are going to start telling me it simply can't be done, when the proof that it can is staring them in the face, then I guess I'll just count myself lucky that I achieved what I've achieved and stick with what I've got. -
Can this plug-in opens jpeg better?
This plug-in claims it can open jpeg file and render it better than photoshop:
http://www.betterjpeg.com/jpeg-plug-in.htm (scroll to the end of the page)
I'm surprised to see this claim. I always assume a jpeg file is rendered the same way by all apps.
Is their claim true? If so, why wouldn't adobe implement this as well?Has a 30 day free trial. Download it, test it and let us know.
-
Power Mac G4 M5183-Can this computer be rebuilt with better hardware?
I recently found my old Power Mac up in the attic, I was courious if I could rebuild the desktop or is the case of this G4 only compatable with its specific hardware. Will the G4 Quicksilver hardware work? Is it even worth trying to rebuild?
With enough money and enough time you can rebuild and update anything. The question is: "Is it worth it?" which only you can answer.
In my experience, the Quicksilver and Digital Audio can be reasonably contemporary and up-to-date performers. Be prepared, however to keep a back up (working) machine and a parts machine available if you intend to rely on these oldsters, and keep a watchful eye and a few spare dollars handy as well.
Older than this - in my view - starts to come under the heading of 'hobby' and can easily become a money pit.
You've been warned.... now go for it!
Jon -
I cannot install Google Toolbar because it's not compatible with Firefox 5. If so I would like to reinstall the previous version of Firefox.
Hey thanx for the help, it worked like a charm.
I think firefox changed the application lay out, as I used to be able to choose my PDF opener from the list at " Adobe Acrobat Forms Document' where I could choose Nitro-PDF reader, but it has now moved all the way to the bottom like you suggested to the obvious place of "Portable Document Format" section.
If someone didn't tell me , I wouldn't have scrolled all the way down to be able to find it... it really should be with all its cousins up the top...
Cheers again.. -
Corrupt Time Portion of a DATE column - how can this happen?
DUMP(START_TIME)
gives
'Typ=12 Len=7: 100,99,255,255,0,0,0'
How can this happen? Is there any way to prevent the database from allowing invalid dates to be inserted.
Thanks in advance,
CadeThat's part of the problem. We don't have direct access to the database. I don't know what the equivalent of SQL Server's Query Analyzer or SSMS in Oracle. I'm more of a 'NET and SQL Server developer.
What the group I am helping does is extract the data (from a system they don't control) using Focus. They send a SQL statement and then Focus retrieves the rowset and does what Focus calls a TABLE FILE operation to put it in a Focus file. Because the Focus rowset typing appears to be based on what datatypes Oracle represents the columns as having, Focus then attempts to interpret the date and time as valid. But apparently they are so messed up that Focus actually halts/crashes.
So every morning, they notify the group responsible for the Oracle system to "fix" the records that are bad. They apparently do this by copying the END_TIME to the START_TIME, hopefully with a simple UPDATE query, but perhaps manually with some other tool.
I'm also not a Focus expert, but in experimenting with sending different SQL statements (like using to_char() to identify the bad data) we did eventually find out how to identify these records using DUMP. What I helped the MIS group here do was to change their SQL query to be:
SELECT CASE WHEN DUMP(START_TIME) = 'Typ=12 Len=7: 100,99,255,255,0,0,0' THEN END_TIME ELSE START_TIME END AS START_TIME
FROM etc. -
Callouts and anchored objects - there must be a better way to do this
I've spent a lot of time in the last six months rebuilding PDF files in InDesign. It's part of my ordinary responsibilities, but I'm doing a lot more of it for some reason. Because I'm sending the text of these rebuild documents out for translation, I like to keep all of the text in a single story. It really helps to have the text in "logical order," I think; when I'm prepping a trifold brochure, I try pretty hard to make sure that the order in which the readers will read the text is duplicated in the flow of the story throughout the ID document.
So, I'm rebuilding a manual that has a 3-column format on lettersize paper, and it's full of callouts. Chock full of 'em. They're not pull quotes, either; each of these things has unique text. Keeping in mind that I'd like the text in these callouts to remain in the same position in the text once I've linked all the stories and exported an RTF for translation, what's the best way to handle them? What I've been doing is inserting an emptly stroked frame as an anchored object, sized and positioned to sit above the text that is supposed to be called out. When my translations come back, they're always longer than the source document, so as I crawl through the text, I resize the anchored frames to match the size and position of the newly expanded translated text, and then nudge them into place with the keyboard.
There Has To Be a Better Way.
There is a better way, right? I'm not actually too sure. If I want to actually fill those anchored frames with text, I can't thread them into the story. I suppose that I could just thread the callout frames and assign two RTFs for translation instead of one, but then the "logical order" of my text is thrown out the window. So, I'm down to asking myself "what's more important? reduction of formatting time or maintenance of the flow of the story?" If there's something I'm missing that would let me dodge this decision, I'd love to hear about it. The only thing I can think of would work like this:
1) Duplicate callout text in the story with a custom swatch "Invisible"
2) Create "CalloutText" parastyle with "Invisible" swatch and apply it to callout text
3) Insert anchor for anchored frame immediately before the CalloutText content
4) Send it out for translation
5) While I'm waiting for it to come back, write a script that would (dunno if this is possible):
a) Step through the main story looking for any instance of CalloutText
b) Copy one continguous instance of that style to the clipboard
c) Look back in the story for the first anchor preceeding the instance of CalloutText
d) Fill the anchored object with the text from the clipboard (this is where I'm really clueless)
e) Apply a new parastyle to the text in the callout
f) Continue stepping through the story looking for further instances of CalloutText
If this really is the only decent solution, I'll just head over to the Scripting forum for some help with d). Can any of you make other suggestions?In-Tools.com wrote:
The use of Side Heads saves weeks of manual labor.
Yup, Harbs, that is exactly what I was describing. If I use the Side Heads plugin to set up a job, will my clients get a missing plug-in warning when they open up the INDD? Will roundtripping through INX strip the plugin but leave the text in the callout? (My clients don't care if the logical flow of the story is broken; it's just me.)
I'm just curious; seems like a pretty obvious purchase to me. I'll probably try to script a solution anyways, after I buy the plugin; that way I get to learn about handling anchored objects in scripts AND deliver the job on time! -
There's got to be a better way to do this (RAM preview frustration)
I loaded a 1:20 second Full HD clip into after effects. I need to edit the video based on certain sounds in the video and see if I'm matching them up correctly by previewing it with sound.
The problem is i'm getting frustrated due to After effects not behaving like Premiere. First who thought it was a good idea not to incorporate sound into after effects? Second, I have an i7 sandy bridge processor, and 16 gbs of ram, yet it still takes time to render the ram preview (with no effects on it yet).
So ram preview is my only option for sound, but the problem is every time I hit ram preview it starts the video all the way from the beginning. This is frustrating as I want to start at a specific point. Imagine having a longer video where the editing needs to take place at the end.
There are people out there doing a lot more complicated professional projects, what do you guys do to get around this?
Why can't after effects do some basic things like premiere like render fast with sound? Is it due to Mercury engine and 64 bits?
This is one of the best products on the market, surely there is a better way to do this right?but the problem is every time I hit ram preview it starts the video all the way from the beginning.
Window --> Preview, enable the "From current time" option
yet it still takes time to render the ram preview (with no effects on it yet).
There is no magic button. If it is compressed, naturally it needs to be decompressed and decoded first. This can consume resources even on fast machines. Furthermore, drive speed matters a lot in such cases. This might actualyl multiply, if you use multiprocessing, so for this kind of simple setup it's usually better to not use it. If your harddrives are fragmented or simply generally slow, frames cannot be loaded as fast and neither will AE be able to use the disk cache. Ergo, convert the footage and move it to the fastest drive in your system.
what do you guys do to get around this?
We preview at reduced compo resolution to extend RAM previews and place markers while the RAM preview plays using the * key on the numpad.
Mylenium -
Better Way To Do This? Selector Operator...
I'm currently writing the selection operator for the algorithm. The aim of it is
to rate how the coursework block have been allocated and give there allocation
a rating...
How I have done it is have a method that searches through the one of the parent
timetables. It looks for coursework time blocks. Once it finds one it notes
this and looks at the next block along. If this is a coursework time block it
notes this as well. I then perform an operation comparing these two coursework
time blocks to find out if they are for the same Module. If they are not this
is not a very effective coursework timetable strategy.
Because of this I note in an array the the position of these two coursework time
block and give them a fitness rating of 1000. I then go on to see if the next
block is a coursework time block. If it is and its not of the same Module ID of
the two previous then I not these 3 block down in an array and give them a
fitness rating of 2000.
My concern is that I am using allot of if and for loop's and the code is
starting to look untidy at best. Is there any better way of doing this?
Below Is my code:
* @param parentOne
* @param parentTwo
public void selectionOperator(ArrayList parentOne, ArrayList parentTwo){
// Store's the fitness rating of sections of the timetable...
ArrayList parentOneBlockFitnessRating = new ArrayList ();
ArrayList parentTwoBlockFitnessRating = new ArrayList ();
//Loops through the timetable's timeblocks...
for (int i = 0; i < parentOne.size(); i++) {
//Checks to see if the current time block is of the class: CourseworkTimeBlock,
//if so it enters this statement...
if(parentOne.get(i).getClass().toString().equals("class Timetable.CourseworkTimeBlock")){
//A temp store for the current CourseworkTimeBlock...
CourseworkTimeBlock tempBlockOne = (CourseworkTimeBlock) parentOne.get(i);
System.out.println("Got Here!, Module ID...: " + tempBlockOne.getModuleId());
//Checks to see if the next time block along is of the class: CourseworkTimeBlock,
//if so it enters this statement...
if(parentOne.get(i+1).getClass().toString().equals("class Timetable.CourseworkTimeBlock")){
//A temp store for the next CourseworkTimeBlock...
CourseworkTimeBlock tempBlockTwo = (CourseworkTimeBlock) parentOne.get(i+1);
System.out.println("Got Here Aswell!");
//Checks to see if the current and next CourseworkTimeBlock module Id's
//are the same, if there arn't then this section is entered...
if(!tempBlockOne.getModuleId().equals(tempBlockTwo.getModuleId())){
//Checks to see if the second time block along is of the class: CourseworkTimeBlock,
//if so it enters this statement...
if(parentTwo.get(i+2).getClass().toString().equals("class Timetable.CourseworkTimeBlock")){
//A temp store for the second CourseworkTimeBlock along...
CourseworkTimeBlock tempBlockThree = (CourseworkTimeBlock) parentTwo.get(i+2);
//Checks to see if all 3 of the Module Id's of the CourseworkTimeBlocks match,
//if they don't match this statement is entered...
if(! tempBlockOne.getModuleId().equals(tempBlockTwo.getModuleId()) && (tempBlockTwo.getModuleId().equals(tempBlockThree.getModuleId()))) {
//ArrayList to store the fitness rating of the current block
//selection...
ArrayList <Integer> blockFitness = new ArrayList<Integer>();
//Position of first block.
blockFitness.add(i);
//Position of second block.
blockFitness.add(i+1);
//Position of second block.
blockFitness.add(i+2);
//Fitness Value
blockFitness.add(2000);
//Add block rating to main rating ArrayList...
parentOneBlockFitnessRating.add(blockFitness);
else{
//ArrayList to store the fitness rating of the current block
//selection...
ArrayList <Integer> blockFitness = new ArrayList<Integer>();
//Position of first block.
blockFitness.add(i);
//Position of second block.
blockFitness.add(i+1);
//Fitness Value
blockFitness.add(1000);
//Add block rating to main rating ArrayList...
parentOneBlockFitnessRating.add(blockFitness);
for (int o = 0; o < parentTwo.size(); o++) {
if(parentTwo.get(o).getClass().toString().equals("class Timetable.CourseworkTimeBlock")){
CourseworkTimeBlock tempBlockOne = (CourseworkTimeBlock) parentTwo.get(o);
System.out.println("Got Here!, Module ID...: " + tempBlockOne.getModuleId());
if(parentTwo.get(o+1).getClass().toString().equals("class Timetable.CourseworkTimeBlock")){
CourseworkTimeBlock tempBlockTwo = (CourseworkTimeBlock) parentTwo.get(o+1);
System.out.println("Got Here Aswell!");
if(!tempBlockOne.getModuleId().equals(tempBlockTwo.getModuleId())){
if(parentTwo.get(o+2).getClass().toString().equals("class Timetable.CourseworkTimeBlock")){
CourseworkTimeBlock tempBlockThree = (CourseworkTimeBlock) parentTwo.get(o+2);
if(! tempBlockOne.getModuleId().equals(tempBlockTwo.getModuleId()) && (tempBlockTwo.getModuleId().equals(tempBlockThree.getModuleId()))) {
ArrayList <Integer> blockFitness = new ArrayList<Integer>();
//Position of first block.
blockFitness.add(o);
//Position of second block.
blockFitness.add(o+1);
//Position of second block.
blockFitness.add(o+2);
//Fitness Value
blockFitness.add(2000);
//Add block rating to main rating ArrayList...
parentTwoBlockFitnessRating.add(blockFitness);
else{
ArrayList <Integer> blockFitness = new ArrayList<Integer>();
//Position of first block.
blockFitness.add(o);
//Position of second block.
blockFitness.add(o+1);
//Fitness Value
blockFitness.add(1000);
//Add block rating to main rating ArrayList...
parentTwoBlockFitnessRating.add(blockFitness);
}As you can see there are allot if statements and some bad coding practice to boot. But I don't know what other ways to do it....
Any directions of other ways how to do this?
Many Thanks
ChrisUnfortunately, I think you're stuck with a bunch of if-statements.
Fortunately, I have some things that may help you.
First, I usually make sure something is of x class via this
if (someObject instanceof SomeClass) {So, I'd adjust your 'class checking' conditionals from this
if(parentOne.get(i).getClass().toString().equals("class Timetable.CourseworkTimeBlock"))to this
if (parentOne.get(i) instanceof Timetable.CourseworkTimeBlock) {Secondly, your code logic is kind of confusing.
Why do you have a for-loop that iterates through every CourseworkTimeBlock, if you then (within each possible iteration) check iteration+1 and iteration+2? What happens if those throw an ArrayIndexOutOfBoundsException, or are null? -
I know there is better ways to do this so that I can merely have it loop infinitely until it hits the max password. I would also like to keep the spot where it has the print Statements because I tend to set some other code in front of it.
If anyone also has some Ideas so I can have it stop and start in particular positions.
Like lets say ?!!!!? to ?~~~~?
Thanks a ton. Below is the code.
char[] password = new char[20];
int current_last_letter = 0;
int max_password = 3;
while(current_last_letter < max_password)
for(int counter_gamma = 32; counter_gamma <= 126; counter_gamma++)
for(int counter_beta = 32; counter_beta <= 126; counter_beta++)
for(int counter_alpha = 32; counter_alpha <= 126; counter_alpha++) //alpha = increase last character by one
password[current_last_letter] = (char)counter_alpha;
System.out.print(password);
System.out.println("::null");
if(current_last_letter > 0)
password[current_last_letter-1] = (char)counter_beta;
else
counter_beta = 127;
if(current_last_letter > 1)
password[current_last_letter-2] = (char)counter_gamma;
else
counter_gamma = 127;
current_last_letter++;
If I interpreted your code right, you simply want to count in radix '~'-' '+1 (read this carefully).
Given a password, and a lo/hi value for the character range, the following method returns the
lexicographically next password. The method returns null if no such password exists -- char[] nextPassword(char[] pw, char lo, char hi) {
for (int i= pw.lenght; --i >= 0;)
if (pw[i] < hi) { // next character possible here?
pw[i]++;
return pw;
else // * no, reset to first value and continue;
pw= lo;
return null; // all pw-chars reached their hi value
So, if you initialize a password char array as follows -- char[] pw= new char[20];
Arrays.fill(pw, ' '); the following fragment handles every possible password -- do {
handlePassword(pw); // do something with pw
pw= nextPassword(pw, ' ', '~');
} while (pw != null); kind regards,
Jos -
When selecting into a variable, is there a better way to do this? I want to make sure I don't get an error if there are no rows found. I think it's inefficient to count the number of occurrences before selecting into the variable. Can I use %NOROWSFOUND or something similar? If so, can you please provide an example? Thanks, Margaret
Tzi := null;
SELECT count(BUSINESS_tz)
INTO Tzi_cnt
FROM BUSINESS O
WHERE ID = Fac_id AND BUSINESS_type = 'Store' AND O.Business_seq = (SELECT Business_seq
FROM Org_lookup
WHERE Store_seq = 0);
IF Tzi_cnt > 0
THEN
SELECT BUSINESS_tz
INTO Tzi
FROM BUSINESS O
WHERE ID = Fac_id AND BUSINESS_type = 'Store' AND O.Business_seq = (SELECT Business_seqthe proper way to do this is to use an exception block;
so
begin
Tzi := null;
SELECT BUSINESS_tz
INTO Tzi_cnt
FROM BUSINESS O
WHERE ID = Fac_id AND BUSINESS_type = 'Store' AND O.Business_seq = (SELECT Business_seq
FROM Org_lookup
WHERE Store_seq = 0);
exception
when no_data_found
then
null; -- put any code to execute here or raise exception
end;you can also use sql%rowcount to find out how many rows your sql statement affected:
eg directly after your sql and before any rollback or commit
declare
v_rowcount number;
begin
insert into <table_name>
select * from <other_table>;
v_rowcount := sql%rowcount;
end; -
Is there a better way to do this projection/aggregate query?
Hi,
Summary:
Can anyone offer advice on how best to use JDO to perform
projection/aggregate queries? Is there a better way of doing what is
described below?
Details:
The web application I'm developing includes a GUI for ad-hoc reports on
JDO's. Unlike 3rd party tools that go straight to the database we can
implement business rules that restrict access to objects (by adding extra
predicates) and provide extra calculated fields (by adding extra get methods
to our JDO's - no expression language yet). We're pleased with the results
so far.
Now I want to make it produce reports with aggregates and projections
without instantiating JDO instances. Here is an example of the sort of thing
I want it to be capable of doing:
Each asset has one associated t.description and zero or one associated
d.description.
For every distinct combination of t.description and d.description (skip
those for which there are no assets)
calculate some aggregates over all the assets with these values.
and here it is in SQL:
select t.description type, d.description description, count(*) count,
sum(a.purch_price) sumPurchPrice
from assets a
left outer join asset_descriptions d
on a.adesc_no = d.adesc_no,
asset_types t
where a.atype_no = t.atype_no
group by t.description, d.description
order by t.description, d.description
it takes <100ms to produce 5300 rows from 83000 assets.
The nearest I have managed with JDO is (pseodo code):
perform projection query to get t.description, d.description for every asset
loop on results
if this is first time we've had this combination of t.description,
d.description
perform aggregate query to get aggregates for this combination
The java code is below. It takes about 16000ms (with debug/trace logging
off, c.f. 100ms for SQL).
If the inner query is commented out it takes about 1600ms (so the inner
query is responsible for 9/10ths of the elapsed time).
Timings exclude startup overheads like PersistenceManagerFactory creation
and checking the meta data against the database (by looping 5 times and
averaging only the last 4) but include PersistenceManager creation (which
happens inside the loop).
It would be too big a job for us to directly generate SQL from our generic
ad-hoc report GUI, so that is not really an option.
KodoQuery q1 = (KodoQuery) pm.newQuery(Asset.class);
q1.setResult(
"assetType.description, assetDescription.description");
q1.setOrdering(
"assetType.description ascending,
assetDescription.description ascending");
KodoQuery q2 = (KodoQuery) pm.newQuery(Asset.class);
q2.setResult("count(purchPrice), sum(purchPrice)");
q2.declareParameters(
"String myAssetType, String myAssetDescription");
q2.setFilter(
"assetType.description == myAssetType &&
assetDescription.description == myAssetDescription");
q2.compile();
Collection results = (Collection) q1.execute();
Set distinct = new HashSet();
for (Iterator i = results.iterator(); i.hasNext();) {
Object[] cols = (Object[]) i.next();
String assetType = (String) cols[0];
String assetDescription = (String) cols[1];
String type_description =
assetDescription != null
? assetType + "~" + assetDescription
: assetType;
if (distinct.add(type_description)) {
Object[] cols2 =
(Object[]) q2.execute(assetType,
assetDescription);
// System.out.println(
// "type "
// + assetType
// + ", description "
// + assetDescription
// + ", count "
// + cols2[0]
// + ", sum "
// + cols2[1]);
q2.closeAll();
q1.closeAll();Neil,
It sounds like the problem that you're running into is that Kodo doesn't
yet support the JDO2 grouping constructs, so you're doing your own
grouping in the Java code. Is that accurate?
We do plan on adding direct grouping support to our aggregate/projection
capabilities in the near future, but as you've noticed, those
capabilities are not there yet.
-Patrick
Neil Bacon wrote:
Hi,
Summary:
Can anyone offer advice on how best to use JDO to perform
projection/aggregate queries? Is there a better way of doing what is
described below?
Details:
The web application I'm developing includes a GUI for ad-hoc reports on
JDO's. Unlike 3rd party tools that go straight to the database we can
implement business rules that restrict access to objects (by adding extra
predicates) and provide extra calculated fields (by adding extra get methods
to our JDO's - no expression language yet). We're pleased with the results
so far.
Now I want to make it produce reports with aggregates and projections
without instantiating JDO instances. Here is an example of the sort of thing
I want it to be capable of doing:
Each asset has one associated t.description and zero or one associated
d.description.
For every distinct combination of t.description and d.description (skip
those for which there are no assets)
calculate some aggregates over all the assets with these values.
and here it is in SQL:
select t.description type, d.description description, count(*) count,
sum(a.purch_price) sumPurchPrice
from assets a
left outer join asset_descriptions d
on a.adesc_no = d.adesc_no,
asset_types t
where a.atype_no = t.atype_no
group by t.description, d.description
order by t.description, d.description
it takes <100ms to produce 5300 rows from 83000 assets.
The nearest I have managed with JDO is (pseodo code):
perform projection query to get t.description, d.description for every asset
loop on results
if this is first time we've had this combination of t.description,
d.description
perform aggregate query to get aggregates for this combination
The java code is below. It takes about 16000ms (with debug/trace logging
off, c.f. 100ms for SQL).
If the inner query is commented out it takes about 1600ms (so the inner
query is responsible for 9/10ths of the elapsed time).
Timings exclude startup overheads like PersistenceManagerFactory creation
and checking the meta data against the database (by looping 5 times and
averaging only the last 4) but include PersistenceManager creation (which
happens inside the loop).
It would be too big a job for us to directly generate SQL from our generic
ad-hoc report GUI, so that is not really an option.
KodoQuery q1 = (KodoQuery) pm.newQuery(Asset.class);
q1.setResult(
"assetType.description, assetDescription.description");
q1.setOrdering(
"assetType.description ascending,
assetDescription.description ascending");
KodoQuery q2 = (KodoQuery) pm.newQuery(Asset.class);
q2.setResult("count(purchPrice), sum(purchPrice)");
q2.declareParameters(
"String myAssetType, String myAssetDescription");
q2.setFilter(
"assetType.description == myAssetType &&
assetDescription.description == myAssetDescription");
q2.compile();
Collection results = (Collection) q1.execute();
Set distinct = new HashSet();
for (Iterator i = results.iterator(); i.hasNext();) {
Object[] cols = (Object[]) i.next();
String assetType = (String) cols[0];
String assetDescription = (String) cols[1];
String type_description =
assetDescription != null
? assetType + "~" + assetDescription
: assetType;
if (distinct.add(type_description)) {
Object[] cols2 =
(Object[]) q2.execute(assetType,
assetDescription);
// System.out.println(
// "type "
// + assetType
// + ", description "
// + assetDescription
// + ", count "
// + cols2[0]
// + ", sum "
// + cols2[1]);
q2.closeAll();
q1.closeAll(); -
Is there a better way to do this function?
Hi,
I was just wondering if someone can suggest a better way of doing the function I have attached below. Basically, it is counting pulses (falling edges) over a 250ms time period and converting that to speed. There seems to be some cases where it counts lesser pulses for a longer period than 250ms which results in lower speeds, but physically the device doesnt run any slower. Or sometimes speed is higher. This code was written by someone else and I am trying to work it better.
V
P.S. There is an overall master While loop.
I may not be perfect, but I'm all I got!
Attachments:
counter_wait.JPG 87 KBVeeJay wrote:
@ yamaeda what do you mean by coercion dots? Can you explain in simpler terms.? Thanks!
V
coercion dots are those red dots on your arithmatic functions. That means you are dividing data of different types. For instance I32 and U8 would give you a coercion dot. What you will want to do is right click your constants, choose representation->[data type]. Either that, or you can go to the numeric-> conversion pallette (i think) thats where it is and convert numbers using the primitives there.
CLA, LabVIEW Versions 2010-2013 -
Is there a better way to do this with Flash?
I am new to Flash but am slowly teaching myself via Lynda.com etc
I have an image that I have added to a website via a content management system and want to make certain areas of that image into links to other sites.
I found this page that does the kind of thing I want to do, but it appears from looking at the source code that the person who has done this has cut the image up into several sections in order to fit it into a table: http://www3.imperial.ac.uk/staffdevelopment/postdocs1/guidance
Is there a better way to achieve the same kind of effect using Flash by making ares of an image into links and keeping the image as a whole?There are ways to keep the image whole and have portions of it linking to different places both in HTML and in Flash. In HTML you can use an image map. In Flash, you can just lay invisible buttons atop the image and use those to link.
Maybe you are looking for
-
Is there a practical limit to the number of airport extreme base stations and airport express units one should have in a home network?
-
New User Database schema and table name
When i create a new user in Oracle Webcenter Spaces 11g, I am not able to get the name of the database schema and table, where it is stored. Any insight on this will be very helpful.
-
Why doesn't siri work away from home
why doesn't siri work way from home
-
Graduating scale and condition type pr002
hi, I am using the pricing condition pr002 in a sales order.the scales i have maintained is between 0-4 items price shd be 10000 and between 4 to 8 price is 8000.however in sales order with basic price ,discount and freight at first ths conditio
-
I am preparing my book for print but all my jpg images that are supposed to be black and white have color in them when I look at the separations preview. Is there an easy way to remove all color from my jpg images?