Vou mostrar aqui uma solução que tenho utilizado frequentemente em alguns trabalhos, substituindo um pouco uma série de tratamentos dentro da query e consequentemente, economizando um tempinho!

Problema:
Como fazer buscas em banco de dados ignorando letras maiusculas/minusculas?

Geralmente os SGBDs são case sensitive e é muito comum recorrermos a alternativas para transformar os textos na hora de comparar valores do banco case sensitive em consultas que fazemos.

Exemplo de uma query comum:

SELECT * FROM cliente WHERE nome LIKE ‘%Marcelo%’;

Nessa query, retornarão apenas clientes que possuam Marcelo (com M maiúsculo) no nome. Se o valor consultado estivesse em letras minúsculas (marcelo), ignoraria todos os nomes em maiúsculos cadastrados no banco.

Alternativa Padrão:

Algumas pessoas passaram a utilizar a função UPPER para eliminar esse problema:

SELECT * FROM cliente WHERE UPPER(nome) LIKE ‘%’ || UPPER(‘Marcelo’) || ‘%’;

Nesse caso, o valor armazenado na coluna nome será transformado para CAIXA ALTA e o mesmo acontecerá para o valor consultado. No caso, o valor Marcelo será transformado para MARCELO.

Isso funciona até certo ponto, pois essa função não converte letras acentuadas e caracteres especiais. No caso, João seria transformado para JOãO. Creio que esse não é o nosso objetivo, nesse caso.

Eis a solução para o problema:

Sendo assim, a melhor alternativa para esse caso seria usar o termo “~*” (til + asterístico). Nesse caso, o Postgre automaticamente procura por qualquer ocorrência do valor pesquisado, ignorando o case sensitive.

SELECT * FROM cliente WHERE nome ~* ‘Marcelo’;

Essa técnica funciona em PostgreSQL e MySQL mas nunca testei no Oracle ou MSSQL. Quem experimentar pode compartilhar conosco o resultado com comentários.

Pesquise por hospedagens de sites com MySQL e PostgreSQL

I hate craving for something.
Like how much I crave for murni/william’s and that bubble tea in Sunway,
that every time I accidentally miss the chance,
I feel so sad and betrayed,
that I can’t help but cry.

Then thinking about it,
I cried for some bloody food and a bloody bubble tea.
It’s so what the heck, even to me.
Yet, I truly, honestly, really cannot help it.

Gosh, I just realize I haven’t seen my friends the whole holidays.
In fact, the last time I’ve seen them was ages ago.
I am devastated.

I bumped into a weird issue that my java web application is working perfectly in Windows box where as when I deploy that in linux it throws following error “dbuser1.table1″ doesn’t exist.

After scratching our head for couple of hours I found that I created table with name as “Table1″ in Mysql database where as in the code I called it as “table1″ and that’s the problem. The weird thing is that it everything works fine if the database is deployed in Windows and all this problem will arise only with Linux.

After that I changed the code to access the table with proper name everything works fine. As I used this table name in more than one place it took more than 30 mins to search and replace all entries. So, whenever you have a possibility of linux deployment make sure you use a DBObjectConstants class for holding all database related object names and use that constant through out your application. In this way fixing these kind of issues will be more easier.

Nothing rhymes with ‘orange’.

But when it comes to search there is an altogether different problem with this word (and many others).

If you’re searching for the colour orange, it’s unlikely that references to the company of the same name, will be relevant to you.

And if that’s the case, wouldn’t it be grand if your search engine took steps to differentiate colour from company?  Afterall, the latest Find- option in IE8 allows you to do it on individual pages.

Providing surfers with a means of case-sensitive searching is one such option – and it used to be available in some free engines (notably Alta Vista).  Today though, only semantic search engines factor in case-sensitivity in search – which is understandable, given they *understand* the difference between orange the colour (and fruit), and Orange the company.

I suppose it could fairly be argued that some aspects of semantic search should obviate the problem of case-sensitivity in any search, but given how far away a reliable, general purpose semantic engine seems at present, wouldn’t it make sense to incorporate this option into the Googles, Bings and Yahoos of this world?

While there has previously been some confusion about case-sensitivity in Google results, (since resolved) it is nonetheless possible to try case-sensitive searching in Google via this bespoke appengine.

That said, there is an important caveat  – those titles or sentences which start with the word Orange are (of course) capitalised, so it can be a bit of a fuzzy way to do things.

Still, it’s an interesting project, and well worth a look.

Descrição: The server encountered an internal error () that prevented it from fulfilling this request.
Exceção:
org.apache.jasper.JasperException: Can’t find bundle for base name “xxxxxx“
java.util.MissingResourceException: Can’t find bundle for base name “xxxxxx“
java.util.ResourceBundle.throwMissingResourceException
java.util.ResourceBundle.getBundleImpl
com.sun.faces.taglib.jsf_core.LoadBundleTag.doStartTag
org.apache.jsp.pages.menu_jsp._jspx_meth_f_loadBundle

O que é?
“Erro 500, falha interna do servidor.
Este erro ocorre sempre que o servidor não consegue interpretar algo estranho no código, todos os códigos de erro que começam com 5 (500, 501, 502, etç.) significam basicamente isto, falha ao tentar entender o código, porém o erro 500 é genérico, o servidor não conseguir realizar a solicitação do código, mas também não sabe por que, os demais erros cujo prefixo é 5, o servidor ao menos sabe por que ocorreu o erro.”

Como resolver?
Repare que as linhas da exeção indicam todas as mesmas coisas:
Can’t find bundle for base name ou throwMissingResourceException, trazendo isso para o português ficaria mais ou menos “Não foi possível encontrar o pacote que se chama XXXXX“, ou seja, nosso pacote não foi encontrado, mas que pacote? basta reparar no que vem descrito depois de Can’t find bundle for base name, no caso do exemplo “XXXXXX“, este é o nome que à página está procurando, mas não o encontrou.
Verifique se este arquivo (properties, bean, xml, etç.) realmente existe e está no local indicado, se ele realmente existir, verifique se o nome tanto do arquivo quando do caminho está correto (br.com.projeto.pasta1.pasta2.arquivo, etç.) e não se esqueça que java é case sensitive, minhaClasse é diferente de MinhaClasse.

.camwhore.yellownails.freestuffs.matchingbottles.theverysweetdriver.snowflake.thinkhappyandyouwillbe.

How could I forget to brag about the whole box of Krisy Kremes that we got from Mei Shan’s good conning skills?
And we got it way before Krispy Kreme was open.
Now to try the other flavours and our lives are complete!

What an unlucky day. I bumped into the last person I ever wanna see just when I reached home.
But at least he reminded me how right I am with my current decision.

Today was fun. I manage to take a real tour of KL. Thanks to Teddy who’s a KL-ian.
I like air mata kucing from now onwards. I don’t know why I’ve never tried it previously.
Tiny made me laugh non-stop today. I loveee listening to her!

I’m being more fickle minded lately. So fickle minded that I think I annoyed Papaya. I skipped breakfast and lunch just because I couldn’t figure out what I wanted to eat. I finally ate at 5pm, not because I was hungry though, more of I was afraid I might pengsan.
We ate at The Loaf and not long after I left, they met Tun Dr Mahathir Mohamad and took a picture with him. Sheesh. If only I stayed longer.

Cupcakes from The Loaf. But none of them are mine.

Today, I feel guilty. But I don’t know for what.
Today, I let out a big secret I should have kept.

On a happier note,

I met, hugged, danced for and took a picture with VJ Utt!

Look at my star struck face. And the super obvious two tones on my arms!
Oh right, and there’s Adam C and Jin too!
Ah, thank goodness Azmi asked them to take the picture!

This is another one of those questions, that gets asked at least once a day in the Microsoft SQL Server programming newsgroups. Over the years, different programmers have come up with different solutions, and I’ll be presenting some of those in this article.

A default SQL Server installation is case insensitive, which means that SQL Server will not differentiate between upper and lower case characters/letters. That is, “Oracle” is the same as “ORACLE” or “oracle”. Let’s see what controls this case sensitiveness of SQL Server. In SQL Server 6.5 and 7.0, you can select a ’sort order’, while installing SQL Server. Sort order defines how the characters supported by the code page are compared and evaluated. The default sort order is ‘case insensitive’. If you want SQL Server to perform case sensitive searches, you need to select binary sort order while installing SQL Server (6.5 or 7.0). In these versions (SQL Server 6.5 and 7.0), sort order is defined at the server level, and cannot be manipulated at the database level. If you want to change the sort order of an existing SQL Server 6.5 or 7.0 instance, you must rebuild the master database using rebuildm utility. Not a pleasant thing to do.

Terminology has changed, and now a sort order is called ‘collation’ in SQL Server 2000. Also ANSI SQL:99 uses the word ‘collation’. By default SQL Server 2000 gets installed with case insensitive collation. You must change the collation of the server while installing, if you want case sensitiveness. But the good news is, SQL Server 2000 lets you specify collation at the database as well as column level. That means, you can have a default SQL Server installation with case insensitive collation, but a specific database or a set of columns can have case sensitive collation.

Lets see why one would want to perform case sensitive searches. Classic example is, password comparisons. People use a combination of upper and lower case (mixed case) characters in their passwords, just to make the passwords difficult to guess. But there is no point in doing that, if the database disregards the case. To keep the users happy, and their passwords secure, programmers prefer case sensitive comparisons in this case.

Here is another example: Some people tend to get the MOST out of their database. Yes, from a performance point of view, a binary sort order or case sensitive collation will perform faster, but only at the expense of the usability. Think about a user searching your table for the phrase “fish and chips”, and not being able to find a match, just because, that phrase is actually stored in the database as “Fish and Chips”. If this happens to be a database serving an online catalog, you will lose some sales as well. One way of avoiding this problem would be, to store everything in lower case, and covert user input to lower case before comparing with the table data. Too much hassle, isn’t it? I wouldn’t recommend this approach. Not worth the performance you gain.

Here are some methods you can employ for case sensitive searching in SQL Server:

All the examples in this article, are based on the following table and data. Requirement is to implement case sensitive search for CustID and CustPassword columns:

CREATE TABLE dbo.Customers(	CustID char(8) NOT NULL PRIMARY KEY CLUSTERED,	CustPassword varchar(15) NOT NULL,	CustName varchar(25) NOT NULL,	CustEmail varchar(60) NOT NULL,	CustAddress varchar(100) NOT NULL)GO

INSERT INTO dbo.Customers VALUES('USA00001', 'TheUnbreakable', 'John', 'John.Smith@SomeDom.com', '10932,Bigge Rd., Menlo Park, CA')

INSERT INTO dbo.Customers VALUES('USA00002', 'ABCxyz012789', 'Heather', 'HGraham@SomeDom.com', '18 Broadway Av. San Francisco, CA')

INSERT INTO dbo.Customers VALUES('EUR00001', 'MagicNSparkle', 'Vyas', 'VyasKN@SomeDom.com', '18 Grove Mead, Herts, AL09 7JK, UK')

INSERT INTO dbo.Customers VALUES('ASA00001', 'aAbBcCXyZ', 'Franklin', 'FMayer@SomeDom.com', '22 Main Street, Singapore')

INSERT INTO dbo.Customers VALUES('OZ000001', 'DOWNundah', 'Steve', 'SWaugh@SomeDom.com', '321, Cricket Way, Melbourne, Australia')

--NOTE: I omitted the VALUES list in the above INSERT statements, for simplicity.--This is not a best practice, and you should always include a VALUES list in your INSERT statements.

Consider the following example, that searches the Customers table for a given Customer ID and password. On a server/database with case insensitive collation, this query will disregard case, and will find a row, as long as the values match, no matter what case is used in the input. Since the @CustID ‘usa00001′ matches a CustID of ‘USA00001′ and @CustPassword ‘theunbreakable’ matches the CustPassword of ‘TheUnbreakable’, the following query prints ‘Customer Found!’:

DECLARE @CustID char(8), @CustPassword varchar(15)

SET @CustID = 'usa00001'SET @CustPassword = 'theunbreakable'

IF EXISTS(	SELECT 1	FROM dbo.Customers	WHERE 	CustID = @CustID		AND CustPassword = @CustPassword)BEGIN	PRINT 'Customer Found!'ENDELSEBEGIN	PRINT 'Invalid Customer ID or Password!'ENDGO

Now our goal is to write queries that perform case sensitive comparisons. I will introduce you to some popular techniques, and show you how to make those methods efficient. It is important to note that, if and when you migrate the following queries onto a case sensitive database, it is better to get rid of the additional bits from the WHERE clause, and compare normally.

DECLARE @CustID char(8), @CustPassword varchar(15)

SET @CustID = 'usa00001'SET @CustPassword = 'theunbreakable'

IF EXISTS(	SELECT 1	FROM dbo.Customers	WHERE 	CAST(CustID AS varbinary(8)) = CAST(@CustID AS varbinary(8))		AND CAST(CustPassword AS varbinary(15)) = CAST(@CustPassword AS varbinary(15)))BEGIN	PRINT 'Customer Found!'ENDELSEBEGIN	PRINT 'Invalid Customer ID or Password!'ENDGO

DECLARE @CustID char(8), @CustPassword varchar(15)

SET @CustID = 'usa00001'SET @CustPassword = 'theunbreakable'

IF EXISTS(	SELECT 1	FROM dbo.Customers	WHERE 	CAST(CustID AS varbinary(8)) = CAST(@CustID AS varbinary(8))		AND CAST(CustPassword AS varbinary(15)) = CAST(@CustPassword AS varbinary(15))		AND CustID = @CustID		AND CustPassword = @CustPassword)BEGIN	PRINT 'Customer Found!'ENDELSEBEGIN	PRINT 'Invalid Customer ID or Password!'ENDGO

DECLARE @CustID char(8), @CustPassword varchar(15)

SET @CustID = 'usa00001'SET @CustPassword = 'theunbreakable'

IF EXISTS(	SELECT 1	FROM dbo.Customers	WHERE 	CustID = @CustID COLLATE SQL_Latin1_General_CP1_CS_AS		AND CustPassword = @CustPassword COLLATE SQL_Latin1_General_CP1_CS_AS)BEGIN	PRINT 'Customer Found!'ENDELSEBEGIN	PRINT 'Invalid Customer ID or Password!'ENDGO

DECLARE @CustID char(8), @CustPassword varchar(15)

SET @CustID = 'usa00001'SET @CustPassword = 'theunbreakable'

IF EXISTS(	SELECT 1	FROM dbo.Customers	WHERE 	CustID = @CustID COLLATE SQL_Latin1_General_CP1_CS_AS		AND CustPassword = @CustPassword COLLATE SQL_Latin1_General_CP1_CS_AS		AND CustID = @CustID		AND CustPassword = @CustPassword)BEGIN	PRINT 'Customer Found!'ENDELSEBEGIN	PRINT 'Invalid Customer ID or Password!'ENDGO

DECLARE @CustID char(8), @CustPassword varchar(15)

SET @CustID = 'usa00001'SET @CustPassword = 'theunbreakable'

IF EXISTS(	SELECT 1	FROM dbo.Customers	WHERE 	BINARY_CHECKSUM(CustID) = BINARY_CHECKSUM(@CustID)		AND BINARY_CHECKSUM(CustPassword) = BINARY_CHECKSUM(@CustPassword))BEGIN	PRINT 'Customer Found!'ENDELSEBEGIN	PRINT 'Invalid Customer ID or Password!'ENDGO

DECLARE @CustID char(8), @CustPassword varchar(15)

SET @CustID = 'usa00001'SET @CustPassword = 'theunbreakable'

IF EXISTS(	SELECT 1	FROM dbo.Customers	WHERE 	BINARY_CHECKSUM(CustID) = BINARY_CHECKSUM(@CustID)		AND BINARY_CHECKSUM(CustPassword) = BINARY_CHECKSUM(@CustPassword)		AND CustID = @CustID		AND CustPassword = @CustPassword)BEGIN	PRINT 'Customer Found!'ENDELSEBEGIN	PRINT 'Invalid Customer ID or Password!'ENDGO

SET NOCOUNT ON

DECLARE @i varchar(500)

CREATE TABLE #t (CharValue varchar(500), BinaryChecksum int)

SET @i = 'A'

WHILE @i <> REPLICATE('A', 500)BEGIN	INSERT #t SELECT @i, BINARY_CHECKSUM(@i)	SET @i = @i + 'A'END

SELECT CharValue, COUNT(*) AS 'Times Repeated' FROM #t GROUP BY CharValueSELECT BinaryChecksum, COUNT(*) AS 'Times Repeated' FROM #t GROUP BY BinaryChecksumSELECT	BINARY_CHECKSUM('A') AS [Checksum value for 'A'],	BINARY_CHECKSUM('AAAAAAAAAAAAAAAAA') AS [Checksum value for 'AAAAAAAAAAAAAAAAA']

DROP TABLE #t

/* Since we have a primary key constraint defined on the CustID column, we cannot alter it directly.First we need to drop the constraint. So we need the constraint name. The following procedure willgive you the primary key constraint name */

EXEC sp_help 'dbo.Customers'GO

/* Substitute your primary key constraint name into the following ALTER TABLE command, to drop the constraint */

ALTER TABLE dbo.Customers DROP CONSTRAINT PK__Customers__71F07EBEGO

/* Change the collation of the CustID column */

ALTER TABLE dbo.Customers ALTER COLUMN 	CustID char(8) COLLATE SQL_Latin1_General_CP1_CS_AS NOT NULLGO

/* Add the primary key constraint back in */

ALTER TABLE dbo.Customers ADD CONSTRAINT PK_Customers PRIMARY KEY(CustID)GO

/* Change the collation of the CustPassword column as well */

ALTER TABLE dbo.Customers ALTER COLUMN 	CustPassword varchar(15) COLLATE SQL_Latin1_General_CP1_CS_AS  NOT NULLGO

/* Now try this query and observe the case sensitive behavior */

DECLARE @CustID char(8), @CustPassword varchar(15)

SET @CustID = 'usa00001'SET @CustPassword = 'theunbreakable'

IF EXISTS(	SELECT 1	FROM dbo.Customers	WHERE	CustID = @CustID		AND CustPassword = @CustPassword)BEGIN	PRINT 'Customer Found!'ENDELSEBEGIN	PRINT 'Invalid Customer ID or Password!'ENDGO

/* Adding two computed columns -- one for CustID and the other for CustPassword */

ALTER TABLE dbo.Customers ADD CustID_CS AS (CAST(CustID AS varbinary(8))),CustPassword_CS AS (CustPassword COLLATE SQL_Latin1_General_CP1_CS_AS)GO

/* Create a nonclustered index on the computed column CustID_CS */

CREATE NONCLUSTERED INDEX NC_NI_Customers_CustID ON dbo.Customers(CustID_CS)GO

/* The following example will make use of the index on the computed column to perform case sensitive searches efficiently */

DECLARE @CustID char(8), @CustPassword varchar(15)

SET @CustID = 'usa00001'SET @CustPassword = 'theunbreakable'

IF EXISTS(	SELECT 1	FROM dbo.Customers	WHERE 	CustID_CS = CAST(@CustID AS varbinary(8))		AND CustPassword_CS = @CustPassword)BEGIN	PRINT 'Customer Found!'ENDELSEBEGIN	PRINT 'Invalid Customer ID or Password!'ENDGO

Dim cn As New ADODB.Connection, rs As New ADODB.RecordsetDim CustID As String, CustPassword As String

CustID = "usa00001"CustPassword = "theunbreakable"

cn.Open "Server=(local);Integrated Security=SSPI;Database=tempdb;Provider=SQLOLEDB"

Set rs = cn.Execute("SELECT CustID, CustPassword FROM dbo.Customers WHERE CustId = '" + CustID + "' AND CustPassword = '" + CustPassword + "'")

If CustID = rs("CustID") And CustPassword = rs("CustPassword") Then    MsgBox "Customer Found!"Else    MsgBox "Invalid Customer ID or Password!"End If

Set rs = NothingSet cn = Nothing

Hope you find the above techniques useful. Here is some additional information that might help you deal with collations: The following command shows you your SQL Server’s default collation:

SELECT SERVERPROPERTY(‘Collation’) AS ‘Server Level Collation’

To see your default database collation:

SELECT DATABASEPROPERTYEX(‘Pubs’, ‘Collation’) AS ‘Database Level Collation’

To see column level collations of Customers table:

EXEC sp_help ‘dbo.Customers’

To see server level collation settings in SQL Server 2000 as well as the previous versions:

EXEC sp_helpsort

To a listing of all available collations in SQL Server 2000:

SELECT * FROM ::fn_helpcollations()

For further information about specific collations:

SELECT COLLATIONPROPERTY(‘German_PhoneBook_CI_AS’, ‘CodePage’)

SELECT COLLATIONPROPERTY(‘French_CI_AS’, ‘LCID’)

SELECT COLLATIONPROPERTY(‘Latin1_General_CI_AS’, ‘ComparisonStyle’)

i was trying to use dbms_stats gather the statistics for a table named: “XTREME_EN”.”Financials”, but got this error:

sys@FUN10U09> exec dbms_stats.gather_table_stats(ownname => ‘XTREME_EN’,tabname
=> ‘Financials’);
BEGIN dbms_stats.gather_table_stats(ownname => ‘XTREME_EN’,tabname => ‘Financials’); END;

*
ERROR at line 1:
ORA-20000: Unable to analyze TABLE “XTREME_EN”.”FINANCIALS”, insufficient privil
eges or does not
exist
ORA-06512: at “SYS.DBMS_STATS”, line 13427
ORA-06512: at “SYS.DBMS_STATS”, line 13457
ORA-06512: at line 1

However, if i use traditional analyze command, it works:

sys@FUN10U09> ANALYZE TABLE “XTREME_EN”.”Financials” COMPUTE STATISTICS;

Table analyzed.

Is that because dbms_stats stupider than analyze? Of course not, Oracle is always encouraging you to use “dbms_stats” instead of using “analyze”.

After search on the metalink, i found this doc: “DBMS_STATS Reports ORA-20000 and ORA-06512 On Case Sensitive Object Names”  Doc ID: 343355.1

https://metalink2.oracle.com/metalink/plsql/f?p=130:14:4774819970862237887::::p14_database_id,p14_docid,p14_show_header,p14_show_help,p14_black_frame,p14_font:NOT,343355.1,1,1,1,helvetica

According to this doc, we have to mention the object name in double quotes.This is very similar to how SQL supports mixed cases.

For our case:

sys@FUN10U09> exec dbms_stats.gather_table_stats(ownname => ‘XTREME_EN’,tabname
=> ‘”Financials”‘);

PL/SQL procedure successfully completed.

Works perfectly!

If you have a Windows box to develop web applications and then, for production and quality, uses Unix / Linux box you should know that there are some little things that you should take care of. No need to mention the fact that on Unix / Linux everything is case sensitive and on not on Windows (even knowing that Apache threats everything as case sensitive on Windows) is the number one thing that can break if you are migrating from Windows to Unix / Linux, but there are some functions that might work differently and break if you don’t take a more chirurgical look.

For a smart and small example take a look on the mkdir function from php. All you need to do is enter the path, writing mode (chmode) and if is recursive. On a Windows box if you enter something like mkdir($path,’0777’,true) or mkdir($path,0700,true) it will create the folders in the path but will as well set them as writable, even if you said not to, but on Unix and Linux that would not work.

Take a look on the first function, mkdir ( $path , ‘0777’ , true );

If you set your script to create the folder path like this in a UNIX box, it will be created, but the permissions will be wrong. The second argument for the function is an integer and not a string, therefore the permissions would not be created correctly and if you tried to upload a file to the directory, it would fail on permission denied. On a Windows box this is totally bypassed.

Another thing to consider is the way that Windows creates the path. If you use something like c:\path\to\root it will fail on a UNIX or Linux box. To avoid this you need simply change the back slash for the forward slash (c:/path/to/root). This will work on both platforms’.

Small cares like this can make your life way easier while developing in PHP.

Have fun.

Last year, I said
“The minute 2008 commence, I was already feeling contented. Although 2007 was really good that I wouldn’t change a thing, I think 2008 has much more in store for me.”

This year I feel the complete opposite. I know I’m not going to feel contented when 2009 commence. I don’t know how anything could beat my celebration last year. And 2008 became cruel to me twas’ the end. I wish I could change everything. But the one thing that remains true is that I know 2009 will have much more in store for me. I don’t think it can get any worse than 2008 anyway.

These were my New Year Resolutions for 2008

1. Get straight A1s for SPM.

2. Be a good Interact President.

3. Be a good secretary for Wenn Jinn.

4. Don’t owe birthday presents.

5. Less egoistic!!

6. Choose a career path.

Well, I accomplished 2/6 of it. I don’t think I was a good president and a good secretary. I’m sure I won’t get straight A1s. What was I thinking? And I’m not sure whether I’m less egoistic now. And I don’t think I know how to fix that.

So this year,

1. Get a full scholarship from Taylor’s.

2. Stay out of relationships.

3. Keep in touch with high school and primary school friends.

4. Make exercising a routine

Every year, I’m getting less ambitious. Haha!
I don’t have a solid plan for countdown tomorrow. I don’t want to end up somewhere pathetic. I want to forget how much I want him to be with me. I know I’ll be thinking about it when they start counting down from 10. Cause that very moment last year, my arms were wrapped around him and I genuinely felt happy.

Well, Bon Voyage 2008!

I know this is damn conceited. But let me be.
Go out with me, I need to stop thinking of nonsense.

Added : Here’s a useful tip for those that wants to break up with someone.
Don’t do it during the holidays!
Cause everytime she’s just watching tv or rotting at home, she’ll start thinking about it and it sucks.
Do it when she’s too busy with her life or something. 
Trust me, I know.
I’m like the pro in getting dumped.

I miei test interni con Oracle 11g su Linux 64 bit di cui ho già scritto in due post precedenti (questo e questo) sono proseguiti a singhiozzo. Qualche giorno fa sono riuscito a installare la patchset 11.1.0.7.0 e i risultati sono a dir poco sorprendenti, ho rilanciato più volte le stesse query che in precedenza di davano errori ORA-00600 e ORA-07445 e queste hanno girato senza alcun problema. Ho rilanciato altre query di update che hanno girato per ore senza alcun problema. Quindi sembra la patchset 11.1.0.7  sistemi realmente qualche problemino presenta sulla prima “release” di Oracle 11g.

L’update che mi ero inventato è andato a buon fine ed ha fatto ciò che in effetti mi aspettavo, però non era sufficente, quindi ho utilizzato un’altro update utilizzando le funzioni del package DBMS_RANDOM per generare delle stringhe casuali ed avere dei valori più distribuiti.

Alla fine però per riprodurre il comportamento descritto da Christian Antognini ho dovuto utilizzare due metodi alternativi: hint oppure parametro optimizer_mode=first_rows_100.

SQL> select * from customers where cust_last_name like ‘cudiz%’;

Execution Plan
———————————————————-
Plan hash value: 2008213504

——————————————————————————-
| Id  | Operation         | Name      | Rows  | Bytes | Cost (%CPU)| Time     |
——————————————————————————-
|   0 | SELECT STATEMENT  |           | 11426 |  1015K|  5625   (2)| 00:01:08 |
|*  1 |  TABLE ACCESS FULL| CUSTOMERS | 11426 |  1015K|  5625   (2)| 00:01:08 |
——————————————————————————-

SQL> select * from customers where cust_last_name like ‘cudiz%’;

Execution Plan
———————————————————-
Plan hash value: 3995041806

—————————————————————————————–
| Id  | Operation                   | Name      | Rows  | Bytes | Cost (%CPU)| Time     |
—————————————————————————————–
|   0 | SELECT STATEMENT            |           |   100 |  9100 |   113   (0)| 00:00:02 |
|   1 |  TABLE ACCESS BY INDEX ROWID| CUSTOMERS |   100 |  9100 |   113   (0)| 00:00:02 |
|*  2 |   INDEX RANGE SCAN          | TEST_IDX  |       |       |    13   (0)| 00:00:01 |
—————————————————————————————–

Francamente trovo curioso come negli esempi di Antognini, in cui crea una tabellina minuscola con 10 record, Oracle 11g vada senza indugio per indice, cosa  apparentemente poco sensata. In realtà mentre scrivevo mi è venuto in mente che probabilmente in questo caso Oracle usa l’istogramma per capire subito che il valore usato dalla query non c’è. Avevo il dubbio che oracle facesse più lavoro andando sul ‘indice per poi dover andare anche sulla tabella, ma non essendoci il valore cercato, il passaggio sulla tabella non c’è. In più ho notato che forzando un full scan sulla tabella oracle fa 3 consistent gets, contro l’1 che fa quando va solo sull’indice, quindi ha ragione lui.

Nel mio caso invece i valori distinti sono oltre due milioni (quasi tutti). Questo fa si che l’istogramma, che può avere al massimo 254 bucket, è per forza approssimato. Rimango perplesso però, perchè dal piano che ho riportato sopra sembra che l’ottimizzatore stimi di estrarre 11426 righe, stima si sbagliata in eccesso, ma che significa comunque meno dell’1% dei numero totale di record presenti nella tabella e nonostante ciò l’ottimizzatore preferisce un full scan. La differenza nei tempi si sente, perchè con l’indice la risposta è pressoché immediata.

Si tratta di test approssimativi, infatti non ho mai avuto problemi con le ricerche per Like sull’applicazione sviluppata nella mia azienda, non è escluso però che si presentino, quindi sicuremente questo test mi tornerà utile e lo approfondirò.

I forgot how cleaning up can make you emotional.

So I packed up yet another bag of things the-guy-i-like gave me.
It wasn’t too long ago when I did the exact same thing and pasted a note that says not to open up the bag until I am completely over him.
Perhaps I’m just bad at handling relationships.

At least I finally took the hardest step in getting over him.

So here comes the Science subjects and my one and only Accounts. And we’re finally done with SPM. Can’t believe it’s so fast. Not complaining.:D
We were never a couple, but I guess you can say what I have with him have been called off just last night. And I think if you know me well enough, you’ll know who wanted it. Don’t confront me about this, please.

No matter where you are registering you will find 2 distincts password handling systems:

• Case sensitive password storage
• Case insensitive password storage

If you work in a LAMP enviroment most likely you will use a case sensitive method, since that’s what you are used to (Linux is a case sensitive system) and most likely you have set on your brain that this is the only secure accepted scenario for password handling, but when you start looking for the user side, case sensitive password handlers became an issue.

Talking with some users (none computer savy) I found out that a password with upper case letters, lower case letters, special chars and numbers combination is a little too much to remember.

Considering this and my will of testing out the poll system from WordPress, I would like to you guys to help me out.

I’m pretty strict with security and I would not exchange security for usability in a password system, but I would like to know what you guys think, so…

If you want to seach the record with sensitive case below code might help you!

(For Case Insensitive)
SELECT password
FROM tblMembers
WHERE (password = @pass COLLATE SQL_Latin1_General_CP1_CI_AS)

(For Case Sensitive)
SELECT password
FROM tblMembers
WHERE (password = @pass COLLATE SQL_Latin1_General_CP1_CS_AS)

Note: @pass is the user’s password

Documentum installation owner user id is used for installing Documentum Content Server. On Windows, the installation sets up services for each repository (docbase), connection broker (docbroker), and Java method server. If you run into issues related to the installation owner id (typically authentication related), there are two aspects that can be checked:

1. Each repository service has an associated command-line which specifies the installation owner as an argument. This command-line is visible on service properties when inspected through Windows services. You can edit the command-line through Documentum Server Manager utility. Select the relevant repository and click on the Edit Service button.
   1. Note that this argument overrides the install owner specified in server.ini in case there is a conflict between the two.
2. All the windows services use logon credentials and the Documentum services mentioned above use the installation owner credentials. These can be edited through the standard Windows service properties dialog.

Inspecting Documentum Repository Service

Issues related to the installation owner may arise when it is a domain user (as opposed to a local user) and/or when multiple content servers are being used. Another aspect to remember is that while Windows treats the user id as case-insensitive Documentum treats it as case-sensitive. Repository configuration also creates a user in the repository for the installation owner, using the exact case as entered during the installation process (at login time prior to installation).

The following are bad ideas that have been used extensively in software engineering, mainly because they are “shortcuts” – methods that are not proper or cogent, but are implemented for the sake of convenience or performance.  Sometimes they are implemented when the programmer has decided to have the application “think” for the user. From a Systems Engineering perspective, these are just wrong. As a coworker once stated, “Performance is the responsibility of the hardware, the software needs to be correct.” In addition, the situation where the computer thinks for the user rather than doing what the user tells it to do, should be limited to very few, well thought-out and critical situations, or quite possibly never.

• Cache - Storing and presenting old data because it takes to long too present more accurate and reliable real time data.
• Cookies – Saving personal data for a website on the computer, especially bad for public computers.
• Remember Me on this Computer (saving passwords) – This violates the fundamental reason for passwords.  And the option is often checked by default on many sites (e.g., Hotmail).
• Multitab Browsers - Because of session beans, these won’t work properly – your session information can get shared between multiple tabs.
• Automatic Updates – Apparently people who still implement this have never had the experience of an application searching and installing updates while they were trying to use it.  Here’s an idea – I tell the software when to look for updates.
• Auto Correction – Who’s driving here, I mean really?! That’s why there is a Spell Check button on many applications. I will spell check and approve when I want to. Thank you.  This is not to be confused with Auto-Completion, which I think is great (example of letting computer help).
• Spaces in File or Directory Names – Although Windows allows it, it can still cause problems in several scenerios.  Ironically, even MS-DOS can’t handle it without quotes.
• Registry – This is spaghetti code for operating systems.  If I were wrong, there wouldn’t be so many registry cleaners out there.  Maybe if it were humanly readable / editable, but that’s probably just asking for trouble.
• Case Indecision - Case Sensitive is OK; Case Insensitive is OK, too;  Case Sensitive Sometimes is pure madness!

To check the Default Server Collation of a SQL Server installation, run the following SQL query:

sp_helpsort

You should get a result similar to this:

Note: To check the collation of a particular database, right click on the database in SQL Server Enterprise Manager (2000) or Management Studio (2005) and click Properties. The collation will be displayed in the Collation Name field.