十二年前的剧本

修订记录：

[ 2013-07-16 01:05 AM GMT -5 nwa ] Changed title from 八年前的剧本 to 十二年前的剧本

[ 2009-07-17 02:18 AM GMT -5 nwa ] Added 给泥猴洗澡, 水晶毛毛虫

[ 2009-06-16 07:51 PM GMT -5 nwa ] 大扁片 -> 大扁片儿，Added two sections: 2002年春 and 2002年秋. Moved 一起走，时间太快 to under 2002年春.

[ 2009-05-13 11:50 AM GMT+8 cya] 是KFC里的生日歌。为此咋们两个还特地驱车前往查证过。

[ 2009-05-13 11:53AM GMT+8 cya] 2001.9.22——这是能忘记的日子么？傻瓜猴。

[ 2009-05-13 11:58PM GMT-5 nwa ] Put a space holder for 2002. 七年前 -> 八年前. 2002年初秋 -> 2001年初秋. 麦当劳 -> 肯德基.

2001年初秋，高二，刚分完文理班。

       同桌

       听不太清她说话

       清晨的车站

       水晶毛毛虫

       第一次牵手

       地下通道里紧张的打地鼠

       蹦蹦床和脏了的棉花糖

       九牛二虎的初吻

       手机振动的快乐

       篮帆文学社

2002年春，磨合的火花和热量

       空中电波情人节

       好吃的巧克力

       胡贵贵和米米米

       一起走，时间太快

2002年秋，冬眠前的储备

       电话亭和草坪

       开自行车兜风

       墙角下的冬眠储备会议

       无线遥控车

       深夜里从公园里拨出的电话

2003年春，高考在即，还有高考前的押注。

       老师请家长，早恋，迷途的少年

       水塔下

       高考前的押注

       课间操窗口的小脸

       崩溃

       拖着尾巴的松鼠

       黑色的你

       半夜的电话，心脏起搏器

       植物人

       大扁片儿

2003年秋，快轨站和肯德基里的生日歌。

       新生宿舍

       周末一起走

       仙剑奇侠传

       不得有惊无喜

       肯德基里的生日歌

       快轨站的电梯

       作弊打水卡

       请你吃米线

       阶梯教室里的同桌

2004年春，5路车，火龙果打吊瓶。

       好又多

       大黑山

       火龙果打吊瓶

       不再担心时间

2004年初秋，灯光明亮的小旅馆。

       狡猾的末班车

       清晨惊悚故事

2005年春，徜徉在舒缓的心情里，吃醋也持续很久。

2005年秋，项链坠和尾巴的咒语，三年的誓言。

       给泥猴洗澡

2006年春，两地时间的手表，如切肤之痛的思念。

2006年秋，重叠的早安和晚安，黑松鼠在叫。

2007年夏，大连机场的猴子棒和百合花。

2008年夏，他们不肯让你走。

2012年春，所有的小纸条漂洋过海。

2012年夏，没有回应的生日祝福。松鼠有了新家，猴子难过，开始流浪。

2013年夏，松鼠不再接电话。流浪猴对人生和爱情的领悟。

2014年夏，十年前松鼠送的电子表，还在默默的运转着。猴子流浪到每天都阳光灿烂的加州，冲浪，兜风，弹吉他，一个人的生活很自由。

2014年秋，松鼠有了女儿。猴子唱爱的代价。

2015年冬，猴子给这个剧本划上一个句号。

PEM -> JKS

 openssl x509 -in client.pem -out client.der -outform DER
 keytool -storepass changeit -keystore client.jks -importcert -alias client -file client.der -trustcacerts

By default java will inline the final member variable

When you changed a final String in a java source file and it's called by a method from a jar, the change won't be picked up.

an alternative to memcached - ramdisk

ramdisk makes things even easier.

[sql ]get quarter end date of a given date

/*
Get the quarter end day of the given date if shift is zero
if shift is -1, it gets the previous quarter end day of the given date
if shift is 1, it gets the next quarter end day of the given date
shift can be +/- 1,2,3,...
*/
FUNCTION getQuarterEnd(ai_date IN DATE, shift INTEGER) RETURN DATE IS
BEGIN
return last_day(add_months(trunc(ai_date,'Q'), 2+shift*3));
END getQuarterEnd;

Win32 WMI

Get available memory:
get-wmiobject -query "select AvailableMBytes from Win32_PerfFormattedData_PerfOS_Memory"

NTSDB Performance Challenge

Essentially it's about finding the fastest way to identify a specific time point in an irregular time series.

SQLiteJDBC 5 row/s vs. CSV

Using SqliteJDBC v056, inserting around 3000 records into a SQlite table, no index, no trigger or any thing, just a plain table - and it took over 10 minutes to finish the job. Roughly 5 records per second.

Switched back to pure text file(CSV), 3000 records, it only took a couple of seconds.

Artificial stupidity in Oracle SQL Developer

Oracle SQL Developer automatically generates GROUP BY clause when it senses there's any aggregation functions in select clause. This might be helpful when writing a new SQL, but when it comes to debugging a huge complex PL/SQL, it is really annoying.

Fortunately there's an option to disable it, Tools > Preferences > Code Editor > Completion Insight > Autogenerate GROUP BY clause.

A template script for debuging PL/SQL procedures

When I get used to the convenience of Eclipse JDT, I can't stand for the awkwardness of Oracle PL/SQL and its IDE. Switching between today's technology and technology from 20 years ago is really a pain. So here is a script template to relief the pain a little bit.

SET serveroutput ON;
DECLARE
ai_instanceid NUMBER;
v_return curType;
rec recType;
BEGIN
ai_instanceid := 5000606;
v_return := plsql_pkg.getparameters(ai_instanceid);
LOOP
FETCH v_return INTO rec;
EXIT
WHEN v_return%notfound;
dbms_output.put_line(rec.portfolionm || ' ' || rec.templateid || ' ' || rec.statementenddt);
END LOOP;
CLOSE v_return;
END;

The mystery of two double quotes and join command(unix)

Two double quotes "" will be interpreted differently by join command.

So when comparing two CSV files, make sure avoiding lines like:

""field1"",""field2""

join will not process these fields.

This is not documented in the manual.

Hard coding

Hard coding (also, hard-coding or hardcoding) refers to the software development practice of embedding input or configuration data directly into the source code of a program or other executable object, or fixed formatting of the data, instead of obtaining that data from external sources or generating data or formatting in the program itself with the given input.

-Wikipedia http://en.wikipedia.org/wiki/Hard_coding

Hard coding has been widely known as a bad practice but I've seen in a lot of programs that constants are being misused to resolve "hard coding" issues. For example, in a java class that has SQL queries, people tend to break down the SQL into pieces and define java constants for the table names or even column names. Their reason is, if the table name or column name changed, it will be easy to change in java class - since there's only one constant for that table or column name. It seems making perfect sense. But think it twice, how often does a table or a column change, how often does the code need to be altered than viewed? Needless to say how hard it is to jumping over the variables when reading the code. Also when there's a change needs to be done, how do you navigate to the places? Isn't it a massive search and replace? So what's the difference between replacing it in one place and replacing it in multiple places? No difference, it's automatic search and replace!
Moreover, embedding SQL in java class isn't really hard coding if you think of the table name as a variable name.

Two ways to use a cursor in a SQL query

To use a cursor in a query, there are two possible ways I’ve used so far:

The first one is using the nested table(or varray). The idea is to read the cursor into the nested table and convert the nest table into a normal table in SQL.
Here is an example:

set serveroutput on;

-- These two types has to be global types, otherwise it will give “PLS-00642 local collection types not allowed in SQL statements”
-- Or maybe you can try using “record” instead of “object”, if it works, it will ease the cursor->nested table part(i.e. using “fetch bulk into” instead of the loop).
-- I’ve tried for a few hours and I gave up…
--create or replace type csrRow as object ( user_nm varchar2(20) );
--create or replace type csrTable as table of csrRow;

declare
cursor csr is select user_nm from table1; -- this is a sample cursor
tmpCsrRow table1.user_nm%type;
csrRow1 csrRow;
csrTable1 csrTable := csrTable(); -- initialize the nested table
counter number;
begin

-- Read the cursor into the nested table(csrTable1)
open csr;
loop
fetch csr into tmpCsrRow;
exit when csr%notfound;
csrTable1.extend;
csrTable1(csrTable1.last) := csrRow(tmpCsrRow);
end loop;
close csr;

--Use the nested table in a regular SQL statement
select count(*) into counter from table(cast(csrTable1 as csrTable));
dbms_output.put_line('Line count:' || counter);

end;



Another way is using pipelined table.
The idea is to create a wrapper for the ref cursor, the wrapper is a pipelined table. And in sql, use “select * from table(pipelinedTable)” to get the data.
Example:

create or replace package pipelinedTest as
type refCsrRec is record ( user_nm varchar2(20) );
type refCsrTable is table of refCsrRec;
function wrapper return refCsrTable pipelined;
end pipelinedTest;
/
create or replace package body pipelinedTest as
function wrapper return refCsrTable pipelined is --This wrapper function converts a cursor into a pipelined table
cursor csr is select user_nm from table1;
tmpRec refCsrRec;
begin
open csr;
loop
fetch csr into tmpRec;
exit when csr%notfound;
pipe row(tmpRec);
end loop;
close csr;
return;
end wrapper;
end pipelinedTest;
/
set serveroutput on;

declare
counter number;
begin
select count(*) into counter from table( pipelinedTest.wrapper() );
dbms_output.put_line('Line count:' || counter);
end;

I tried to write the wrapper() function as an inner function, but got some problems(I think it’s about SQL engine can’t see the dynamically defined PL/SQL function) so I just put it in a package like what we did for Cognos.


Conclusion
PL/SQL is short for a Pathetic Language/Seems Quite Lame

Software Development, $ by time makes simple things complex

A software development company shouldn't be like a garage - doing repeating jobs and charge clients on how much time developers have spent on it.

This makes a simple job complex, this impairs efficiency, discourages innovation... I don't think it is necessary to give any examples to proof this. You know what it is like if you've worked in a traditional software company.

Then what should software development look like?
I don't know yet.

Java is a good tool to do prototyping

when it comes to production, it is not fast enough.

Cognos 8 doesn't support ORACLE REF CURSOR

So to get around with this, use wrapper function and PIPELINED table.

JUnit is taken out of NOWING projects

JUnit is a very convenient library. Like an automatic transmission for a family car - You press the paddle, it runs.

But NOWING projects will go back to the previous unit testing method - each class that needs to be tested will have a static main method and the test will be done there.

CAUTION: Oracle treats empty string as null

select 'oracle empty string is equal to null' from dual where '' is null;

Apache htaccess password

1.AccessFileName .htaccess in httpd.conf (if windows)

2.httpd.conf or vhost.conf
<directory "path/to/webfolder>
Options None
AllowOverride all
Order deny,allow
</directory>
3.htpasswd -c -b password.txt name password

4.Create .htaccess under webfolder
AuthUserFile /path/to/passwd.txt
AuthType Basic
AuthName "some text"

<limit GET POST>
require valid-user
</limit>

NOWING Automated Currency Trading System Structure

Hey! Smile!

• Book Center
  stores historical quotes, transaction records and periodical analysis results.

• Broker Gateway
  communicates with FX brokers' systems, sends/receives transactions.

• Automated Traders
  hold different trading strategies and send out trading signals.

• Communication Interface
  controls and monitors the system.

• Web Server
  provides a fancy interface, shows charts, reports to end users.

• End User
  unhappy investors.