Best way to implement oracle TEXT on multiple tables with regular updates

Hi,
I have the following situation:
5 tables where we want full text search on multiple columns.
Some of the tables have a master/detail relation. (1 to 1000, or more)
because of the number of transactions on these tables we can't have a lag in the sync time.
Currently I have create a dummy table just for the search with 2 columns: for the primary key to all the other tables and one for the update trigger.
I use the user_datastore with a procedure to join all the necessary columns resulting in a clob.
My question is regarding the update.
Of course I can create triggers to update the dummy field in the search table, but this will give lot of updates on that table with possible locking issues.
What would be the best approach to have this search functionality working?
I am open for any ideas!
Thanks,
Edward

Ok, I will focus on building a solution on 12c.
right now I have used a USER_DATASTORE with a procedure to glue all the field together in one document.
This works fine for the search.
I have created a dummy table on which the index is created and also has an extra field which contains the key related to all the tables.
So, I have the following tables:
dummy_search
contracts
contract_ref
person_data
nac_data
and some other tables...
the current design is:
the index is on dummy_search.
When we update contracts table a trigger will update dummy_search.
same configuration for the other tables.
Now we see locking issues when having a lot of updates on these tables as the same time.
What is you advice for this situation?
Thanks,
Edward

Similar Messages

Best way to remove duplicates based on multiple tables

Hi,
I have a mechanism which loads flat files into multiple tables (can be up to 6 different tables) using external tables.
Whenever a new file arrives, I need to insert duplicate rows to a side table, but the duplicate rows are to be searched in all 6 tables according to a given set of columns which exist in all of them.
In the SQL Server Version of the same mechanism (which i'm migrating to Oracle) it uses an additional "UNIQUE" table with only 2 columns(Checksum1, Checksum2) which hold the checksum values of 2 different sets of columns per inserted record. when a new file arrives it computes these 2 checksums for every record and look it up in the unique table to avoid searching all the different tables.
We know that working with checksums is not bulletproof but with those sets of fields it seems to work.
My questions are:
should I use the same checksums mechanism? if so, should I use the owa_opt_lock.checksum function to calculate the checksums?
Or should I look for duplicates in all tables one after the other (indexing some of the columns we check for duplicates with)?
Note:
These tables are partitioned with day partitions and can be very large.
Any advice would be welcome.
Thanks.

>
I need to keep duplicate rows in a side table and not load them into table1...table6
>
Does that mean that you don't want ANY row if it has a duplicate on your 6 columns?
Let's say I have six records that have identical values for your 6 columns. One record meets the condition for table1, one for table2 and so on.
Do you want to keep one of these records and put the other 5 in the side table? If so, which one should be kept?
Or do you want all 6 records put in the side table?
You could delete the duplicates from the temp table as the first step. Or better
1. add a new column WHICH_TABLE NUMBER to the temp table
2. update the new column to -1 for records that are dups.
3. update the new column (might be done with one query) to set the table number based on the conditions for each table
4. INSERT INTO TABLE1 SELECT * FROM TEMP_TABLE WHERE WHICH_TABLE = 1
INSERT INTO TABLE6 SELECT * FROM TEMP_TABLE WHERE WHICH_TABLE = 6
When you are done the WHICH_TABLE will be flagged with
1. NULL if a record was not a DUP but was not inserted into any of your tables - possible error record to examine
2. -1 if a record was a DUP
3. 1 - if the record went to table 1 (2 for table 2 and so on)
This 'flag and then select' approach is more performant than deleting records after each select. Especially if the flagging can be done in one pass (full table scan).
See this other thread (or many, many others on the net) from today for how to find and remove duplicates
Best way of removing duplicates

What is the best way of selecting 99 colums from a table with 100 colums?

Hi All,
Suppose in a table T, I have 100 colums .I want to select 99 columns.So what is the best way to do the same?
Anyone Help!
Thanx in advance.
Piyush

Try this
SQL> ed
Wrote file afiedt.buf
1 select column_name ||',
2 ' from user_tab_columns
3 where
4 table_name = 'EMP'
5* and column_name not like 'EMPNO'
SQL> /
COLUMN_NAME||','
ENAME,
JOB,
MGR,
HIREDATE,
SAL,
COMM,
DEPTNO,
7 rows selected.
SQL>

Best way to implement active directory in multiple locations

Hi,
Currently we don't have an active directory domain and looking in to configuring a test setup for it.
We have 6 countries and in some countries we have 2 to 3 sites. There is a constant VPN connection between all the locations.
Our users are travelling between the sites. IT is managed from a central location and have one IT responsible on each site which also have to create / modify users.
Should we go for one domain with a domain controller in each site? Or should we go for a parent DC at central location with child DC (sub domains) at the other sites?
What are the pro's and cons of each scenario?
Kr,
Joeri

Hi jfeyen,
I think you have some misunderstanding about OU and site in AD.
OU is the purpose container that can be used to group most other object classes together for administrative purposes. An organizational unit in Active Directory is analogous to a directory in the file system; it is a container that can hold other objects.
And it represent the logical structure of your organization as domain.
Sites in Active Directory represent the physical structure, or topology, of your network. Active Directory uses topology information, stored as site and site link objects in the directory, to build the most efficient replication topology. A site is a set
of well-connected subnets. Sites differ from domains.
For your information, please refer to the following articles:
Organizational Units
http://technet.microsoft.com/en-us/library/cc978003.aspx
Sites overview
http://technet.microsoft.com/en-us/library/cc782048(v=ws.10).aspx
So if a user or a computer connected to the domain, he will be located in the OU which is configured. And this will not change except the configuration changes.
As for the question If we use site & services our workstations will automatically find the right DC?, please refer to the following article:
Finding a Domain Controller in the Closest Site
http://technet.microsoft.com/en-us/library/cc978016.aspx
Regards,
Lany Zhang

Which is the best way to get data from user difine tables with condition

Hi everyone,
I would like using SBO DIAPI to get UDT data rows, there is a method of UserTable.getbykey can get one row information, but I like to get a list of rows from UDT, and the get list have its condition such as: field1="Type_A" ect.
also I want to avoid using doquery get data from DB, because there is direct access DB potential Suspect.
I am using SBO 2005 PL09 & visual studio(VB.net) 2005 and I only using DIAPI do not using UIAPI.
Thanks for any advice.
Kathy

Hi Kathy,
Unfortunately SBObob.GetObjectKeyBySingleValue only works for the B1 business objects as listed in the documentation.
So, when you don't know the keys (= Code) of the records you want to get from your UDT - there's no way around using Recordset object's DoQuery method!
Sorry,
Frank

Displaying Multiple Values on GUI components - best way to implement

Hi,
my program needs to implement a basic function that most commercial programs use very widely: If the program requires that a GUI component (say a JTextField) needs to display multiple values it either goes <blank> or say something more meaningfull like "multiple values". What is the best way of implementing it?
In particular:
My data is a class called "Student" that among other things has a field for the student name, like: protected String name; and the usual accessor methods (getName, setName) for it.
Assuming that the above data (i.e. Student objects) is stored in a ListModel and the user can select multiple "Students", if a JTextField is required to display the user selection (blank for multiple selections, or the student "name" for a single selection), what is the best (OO) way of implementing it? Is there any design pattern (best practice) for this basic piece of functionality? A crude way is to have the JTextField check and compare all the time the user selections one by one, but I'm sure there must be a more OO/better approach.
Any ideas much appreciated.
Kyri.

Ok, I will focus on building a solution on 12c.
right now I have used a USER_DATASTORE with a procedure to glue all the field together in one document.
This works fine for the search.
I have created a dummy table on which the index is created and also has an extra field which contains the key related to all the tables.
So, I have the following tables:
dummy_search
contracts
contract_ref
person_data
nac_data
and some other tables...
the current design is:
the index is on dummy_search.
When we update contracts table a trigger will update dummy_search.
same configuration for the other tables.
Now we see locking issues when having a lot of updates on these tables as the same time.
What is you advice for this situation?
Thanks,
Edward

Best way to implement a simple read of multiple stings from a db

Hello guys,
I am trying to read a collection of countries (Strings) into a collection inside a Movie object. The countries reside in a separate DB table with a FK referencing the Movie.
My questions:
1) What is the best way to implement this? Do I need a seperate entity 'country' ? Is there a general way to read many strings into a collection?
2) the countries list does not have to be updated, only read. How do I make sure it is not persistent?
I have noticed that in my JPA implementation (in netbeans) I do not have something like CollectionOfElements (that is present in hybernate).
Thank you,
Marek

>
My questions:
1) What is the best way to implement this? Do I need a seperate entity 'country' ? Is there a general way to read many strings into a collection?
Probably yes, you need to have one entity for country.for getting the right answer, Post it here - Enterprise Technologies - Java EE SDK (http://forums.sun.com/forum.jspa?forumID=136)

Best way to implement m to n relation?

Could you please give me some advice on the best way to implement m-n relations in apex?
Example: I want to store server and database info. A server can have multiple database of course. But in case of a RAC database, the database can be running on multiple servers. So I have tables:
create table SERVERS (id number primary key, name varchar2(30));
create table DATABASES(id number primary key, name varchar2(30))
and an m-to-n table
create table SERV_DB(serv_id number references SERVERS(id), db_id number references DATABASES(id), instance_name varchar2(30))
So the table SERV_DB can tell me e.g. that database PROD is running on server 'prdsrv1' with instance PROD1 and on server 'prdsrv2' with instance name PROD2
How would you design an apex page to maintain this information (adding relations, updating instance names, etc)? I have a solution using checkboxes and 2 for-loops over htmldb_application.g_f40 (to process checkboxes) and g_f41 (to process text fields with instance names) and some delete statements but the logic behind it looks too complex for me. I am convinced that this can be done more simpler. Seems like a common problem to me, so I wonder if there is no out-of-the-box solution in apex for this?
Could you please show me or create a small demo application with the solution that looks most elegant to you?
Thanks,
Geert

Thanks for your reply. You modified the question slightly, but conceptually it is still the same. What you call the instances table corresponds with my serv_db table. I understand the solution you propose using the tabular report. If I see it correctly, you would have an insert button above the tabular report for each new relation (instance - server) you would want to add. This is ok for the case i used (databases, servers, instances) because there are relatively few relations. However I would not like this solution for other cases. E.g.:
case: you have a list of persons and a list of tasks. You want to assign tasks to persons in a way that each person has multiple tasks to do and each task can be assigned to multiple persons.
Suppose you add a new person, and you want to assign 15 tasks to him. The solution above (with the tabular report) would be quite some work because you would have to click 15 times on the insert button and on each click select a task from the select list. In this case it would be more appropriate to, after selecting the new person, see a list of all tasks (e.g. 30 tasks) with a checkbox in front, so you can mark 15 out of the 30 checkboxes and press submit. It gets more complex when you want to assign also an attribute to the relation, i.e. showing the list with all tasks, a checkbox in front and a select list next to each task where you can choose from e.g. "priority high" or "priority low" to indicate that this task is high or low priority for that person. Is there an easy way to implement that?

Best way to implement a shared Blocking Queue?

What's the best way to implement a shared Blocking Queue that multiple JVMs can enqueue objects in and multiple JVM's can dequeue from simultaneously?
Also, I see references on the web to com.tangosol.coherence.component.util.queue.ConcurrentQueue but I don't see it in the current API docs...
Thanks

Hi snidely_whiplash,
snidely_whiplash wrote:
What's the best way to implement a shared Blocking Queue that multiple JVMs can enqueue objects in and multiple JVM's can dequeue from simultaneously?
Also, I see references on the web to com.tangosol.coherence.component.util.queue.ConcurrentQueue but I don't see it in the current API docs...
ThanksThat class is an internal class, AFAIK.
As for implementing a queue, you might want to look at Ashish Srivastava's ezMQ component for some ideas:
http://ezsaid.blogspot.com/2009/01/implementing-jms-queue-on-top-of-oracle.html
Best regards,
Robert

Best way to implement a docking container

Hi everyone,
I need to create a docking container that behaves the same way the draggable panel system works in Flex Builder (Eclipse):
- Dock other containers to the edges
- Automatically layout other containers when the dragged container is docked into/out of an edge
Is there a custom control that would enable me to do this? If not, what's the best approach to create one?
Thanks,
Amir

Hi snidely_whiplash,
snidely_whiplash wrote:
What's the best way to implement a shared Blocking Queue that multiple JVMs can enqueue objects in and multiple JVM's can dequeue from simultaneously?
Also, I see references on the web to com.tangosol.coherence.component.util.queue.ConcurrentQueue but I don't see it in the current API docs...
ThanksThat class is an internal class, AFAIK.
As for implementing a queue, you might want to look at Ashish Srivastava's ezMQ component for some ideas:
http://ezsaid.blogspot.com/2009/01/implementing-jms-queue-on-top-of-oracle.html
Best regards,
Robert

What's the best way to detect that text fits into ContainerControllers without scrolling?

Hi.
Question
What's the best way to detect that text typed by user (or added programmatically) exceeds available container space and find out where starts truncated part? Is there available some other (than described bellow) easy way to detect it or disallow set of controllers to receive more characters that can be displayed in given composition area?
My partialy failed attempt (simplified)
For example lets say that I've got an editable textflow with attached two instances of ContainerController.
var flow:TextFlow = createSomeFlowFromGivenString(sampleText),
    firstController = new ContainerController(firstSprite, 100, 30),
    lastController = new ContainerController(secondSprite, 600, 30);
flow.interactionManager = new EditManager(new UndoManager());
flow.flowComposer.addController(firstController);
flow.flowComposer.addController(lastController);
flow.flowComposer.updateAllControllers();
With enabled vertical scroll policy I can compare height of the composition in last controller with height of the content:
var bounds:Rectangle = lastController.getContentBounds(),
    overflow:Boolean = lastController.compositionHeight < bounds.height;
trace('Content does not fit into given area?', overflow)
But when I switch vertical scroll policy off (lastController.verticalScrollPolicy = ScrollPolicy.OFF) - unfortunately this no longer works...(In my case scrolling should be dissabled, since text areas might have only one line with restricted width)
Use case
I want to create fillable form. Field might have a single or multiple lines. One field might start in the middle of the page, continue in the next line where it spreads through whole page and end in the - quarter of page width long - third line. Text typed by the user can't exceed given area since it might cover some static text that sits right after/below field.
Something like ascii image bellow:
|                <PAGE>                    |
|                                          |
|                                          |
|                                          |
|               [Field starts here........ |
| ........................................ |
| ........................................ |
| Ends here..]                             |
|                                          |
|                                          |
| [Another field] xxxx xxxx xxxxxxxx x xx |
| xxxxxxxxxxxxxxxxxxx                      |
|                                          |
|                              [One more.. |
| .....]                                   |
|                                          |
|                                          |
|                                          |
|                                          |
|                                          |
|                                          |
|                                          |
|                                          |
Info:
[......] <-- form fields starts with '[' character, and ends with ']'
xxx       <-- sample, static text
| and _   <-- page borders

I've tried something like that previously, but it didn't work because then I've got vertical scroll policy to set to 'auto'. After changing verticalScrollPolicy to ScrollPolicy.OFF in the last container it worked like a charm.
Thanks zhen bian

Best way to implement app wide process

I am working on an application that has a list of art work in an sql report.. I want to be able to add each piece to a collection of items (which is stored in a table), so I added another column to the sql report to store the id of the art work. I am just wondering what is the best way to implement the process to insert the item to the collection of items. Since I want to do this from multiple pages - link from art work sql report; link from the edit page for a particular art work; and on the artists page where there is also an sql report, I thought it may be best to have an application process, to save duplicating the process on multiple pages.. however I don't think there's a way to pass the id of the item to the application process? as I was thinking best to have the application process conditional based on request, and then have the link as a doSubmit().
So is it best just to have the link redirect to the same page, passing the id of the item, and then having a process that runs before header that inserts the item, then assigning the item to NULL?
Thanks in advance,
Trent

Daniel:
Thanks for the info. I have found that the IFRAME works nicely for some applications,
however, some applications I want to re-write the front end to look better in
the portal rather than screen scrape.
Thanks for the help
Ryan
"Daniel Selman" <[email protected]> wrote:
Ryan,
Check out the thread (a few days old) titled "Different ways of creating
portlets".
Sincerely,
Daniel Selman
"Ryan Richards" <[email protected]> wrote in message
news:3d2eda12$[email protected]..
I have a few existing applications that I need to port over to portletsin
Weblogic
Portal 4.0. One application is a servlet based web application witha few
html front-end
screens. I am trying to determine how to do this in the best way. Ihave
noticed
that portlets behave differently inside the portal than do stand-aloneweb
apps.
Any help would be appreciated.
The other web application is the Microsoft Outlook Web Access. (Thisone
is going
to be difficult because it is actually an ASP app. I dont know if thisis
possible
without building some proxy code between bea and iis.)
Thanks
Ryan

Best way to implement service in 3-tier webarchitecture

Hi,
What would be the best way to implement a service in the following 3-tier TopLink architecture: no ejb and a webclient (jsp/servlets)?
Currently I have a server session type and a service pojo using a singleton SessionFactory. The service pojo is used by the jsp/servlets.
Should the service pojo be a singleton? Should a SessionManager be used?
In the examples on OTN a clientsession is directly acquired in the jsp/servlets.
Thanks,
Ronald

Ronald,
There are numerous ways to do this. I would recommend using the SessionFactory. The SessionFactory makes use of the SessionManager, which holds onto the singleton Server session. You can then acquire client sessions using the oracle.toplink.sessions.Session interface from the SessionFactory. There is no need to hold the SessionFactory in a singleton yourself and you do not need to reference the ClientSession directly either.
We have introduced this latest approach to simplify coding. It will work in both the web architecture with JSP/Servlets accessing the sessions as well as within an EJB tier (Session/Message Driven beans).
Doug

Event Structures-best way to implement this UI?

I am trying to write a VI to control & read data from 4 different "channels" (each measuring a DUT) at once. I have written all the VI's for initializing instruments, communicating with the devices (VISA, GPIB), setting bias, reading data, etc...that has all completed. I just need to write the overall program with the user interface to allow the user to control these 4 channels & display the measured data.....as it turns out, this is the tricky part! My head is spinning from trying to figure out how to handle all the possible events.
Basically for each channel, I want the user to be able to
-enable/disable it for measurement (e.g. if there is no device loaded in Ch.3, we don't want to measure Ch.3..maybe disable/grey everything)
-set bias conditions (only if channel enabled). Allow user to change bias "in real-time" by increment/decrementing (e.g. incrementing from 5.00 V to 5.01 V, for example).
-turn biasing on/off (again, only if channel is enabled)
Also, I want each channel to display its measured data (e.g current, temperature reading)..every second or so. No graphs or anything fancy (for now! ), just numeric indicator.
Honestly, this all sounds so simple but I'm having trouble figuring out the best way to implement this, due to the fact that 1) there are multiple channels needing to be monitored for events 2) large number of user events that could occur (seems like at least 4 per channel - enabling/disabling, turning bias on/off, incrementing/decrementing bias values, etc ), Also the if a channel IS enabled, i want to be continously reading/displaying the data. What is the best way to handle this? Should i have 4 separate while loops, each with an event structure to handle events for that particular channel..or will that give me grief somewhre?
Also, I have another nagging question. Pretty much all the examples I see re: Event Structures and booleans involve latched booleans, eg. buttons that are just pressed once and pop back up...e.g. buttons you press to tell it to complete a task (e.g. "Acquire Data" or "Stop") , and once it's pressed it's over and reset. In my case, some of the booleans would not be latched...e.g. the "Enable Ch.2" button would be 'TRUE" as long as i want Ch. 2 to be read....does that make sense? Then, say hours later, if i did want to disable that channel, i would change it to "FALSE" and while that would be an "value change", the new value would not be "TRUE"..does that make sense? So not sure if that would be dealt with the same way in an Event Structure.
Hope this all makes sense and many thanks in advance for any help!!!

You're halfway there. I'd say the best solution is a producer/consumer structure, the event structure is used to generate queued commands to the consumer loop.
All data is handled in the consumer loop, where you among other things have an array of clusters of channel/instrument settings. (I usually have several cluster, one for test data, one for instrument settings, one for general settings and so on)
The event structure can have a 1 sec timeout, in which you queue up a Measure command.
In the consumer, in the measure state you loop through your instruments and if enabled you measure them and update their indicators.
The general (smart) way to setup the queue is with a cluster containing 2 elements, a typedef'd Command and a variant.
This way you can send data with the command in any form which is then interpreted in the consumer.
If, e.g. you press the Enable button on a channel, you can enqueue Enable and the channel number.
/Y
LabVIEW 8.2 - 2014
"Only dead fish swim downstream" - "My life for Kudos!" - "Dumb people repeat old mistakes - smart ones create new ones."
G# - Free award winning reference based OOP for LV

Best way to implement connection pooling with tomcat4.0?

Hi all!
I need your help:
What's the best way to implement connection pooling (Oracle Database) with tomcat 4.0?
I only found ways to implement it in tomcat 4.1, not in 4.0....
Thanks!
Michael

You can use a Datasource managed by tomcat. Earlier versions of tomcat used Exolab's Tyrex for the implementation of the Datasource API which uses connection pooling. Current version uses commons-dbcp and commons-pool (jakarta projects) I think.
You've got to declare the Datasource in server.xml and then in web.xml of your web app.

Best way to implement oracle TEXT on multiple tables with regular updates

Similar Messages

Maybe you are looking for